#Yazılım​​ #Mühendisliği​​ - Ders 13: Güvenilirlik Mühendisliği, Güvenilir Sistem Mimarileri

[FurkanGozukara]

#Yazılım​​ #Mühendisliği​​ - Ders 13: Güvenilirlik Mühendisliği, Güvenilir Sistem Mimarileri

Full tutorial: https://www.youtube.com/watch?v=2tL0ylDPtgo

#Ücretsiz #Ders #Kurs #Eğitim

İyi bir yazılım programlamak sadece kodlama değil aynı zamanda bir mühendislik sürecidir. Büyük ölçekli gerçek hayat sistemleri, kodlanmaya başlamadan önce iyi planlanmalı ve organize edilmelidir. Böylesine büyük ölçekli bir projede başarıya ulaşmak için, fikirden nihai ürünün teslimatına kadar, projenin tüm aşamalarının iyi bir şekilde dokümente edilmesi ve belirlenen kurallara uyulması gerekir. Bu derste kod yazmaktan ziyade nasıl daha iyi bir yazılım mühendisi olacağınız öğretilecektir.

Bilgisayar Becerileri dersi oynatma listesi : https://www.youtube.com/playlist?list=PL_pbwdIyffsk7Rfb3OPCTnDdfKV3cNl5l

Dersin GitHub Depo Adresi : https://github.com/FurkanGozukara/Yazilim-Muhendisligi-IT522-2021

Dersin Discord kanalı bağlantı adresi : https://discord.gg/6Mrb8MwteQ

Discord nasıl çalışır / kullanılır : https://youtu.be/AEwPtYiLvsQ

Bu dersi almak için herhangi bir ön şart gerekmemektedir.

Eğer programlama öğrenmek istiyorsanız veya kendinizi geliştirmek istiyorsanız aşağıdaki derslerimizi de takip edebilirsiniz:

[1] C# ile Programlamaya Giriş dersi ders videoları oynatma listesi : https://www.youtube.com/playlist?list=PL_pbwdIyffskoSXySh0MdiayPJsBZ7m2o

[2] C# ile İleri Programlama dersi ders videoları oynatma listesi : https://www.youtube.com/playlist?list=PL_pbwdIyffslHaBdS3RUW26RKzSjkl8m4

[3] C# ile Nesne Tabanlı Programlama dersi ders videoları oynatma listesi : https://www.youtube.com/playlist?list=PL_pbwdIyffsnH3XJb66FDIHh1yHwWC26I

[4] C# ile ASP.NET Core MVC tabanlı BootStrap ile responsive web tasarımı : https://www.youtube.com/playlist?list=PL_pbwdIyffsnAWtgk4ja3HN3xgMKF7BOE

[5] Yapay Zeka ve Makine Öğrenmesi (örnek programlar C# dilinde) dersi ders videoları oynatma listesi : https://www.youtube.com/playlist?list=PL_pbwdIyffskVschrADCL6KEnL_nqDtgD

[6] Yazılım Mühendisliği dersi ders videoları oynatma listesi : https://www.youtube.com/playlist?list=PL_pbwdIyffslgxMVyXhnHiSn_EWTvx1G-

[7] Bilgi Sistemlerinin Güvenliği dersi ders videoları oynatma listesi : https://www.youtube.com/playlist?list=PL_pbwdIyffslM_o92NwkaUzD7C6Fekx26

[8] Bilgisayar Becerileri dersi ders videoları oynatma listesi : https://www.youtube.com/playlist?list=PL_pbwdIyffsmyE2e909ea1MXLcMb8MenG

Video Transcription

• 00:00:03 Hello everyone,

• 00:00:08 welcome to our 13th software engineering course today.

• 00:00:13 The topics covered in the software engineering section are as follows: redundancy and diversity.

• 00:00:16 redundancy and diversity. Basic approaches to achieving fault tolerance.

• 00:00:18 Basic approaches to achieving fault tolerance. Reliable processes are given.

• 00:00:21 Reliable processes are given. How processes can be used to perform tasks in systems.

• 00:00:24 How processes can be used to perform tasks in systems. System architectures.

• 00:00:28 System architectures. Architectural models for software fault tolerance

• 00:00:30 Architectural models for software fault tolerance are given.

• 00:00:34 are given. Programming guidelines for preventing errors

• 00:00:38 Programming guidelines for preventing errors were given. Software can be trusted.

• 00:00:41 were given. Software can be trusted. Generally, software customers

• 00:00:44 Generally, software customers generally expect all written

• 00:00:47 generally expect all written software to be secure.

• 00:00:49 software to be secure. If

• 00:00:50 If you're buying software or

• 00:00:53 you're buying software or writing it from scratch, you expect this section to be secure.

• 00:00:55 writing it from scratch, you expect this section to be secure. However, for non-critical applications,

• 00:00:58 However, for non-critical applications, some system errors

• 00:01:00 some system errors may be acceptable. This is quite possible.

• 00:01:03 may be acceptable. This is quite possible. Some applications have

• 00:01:06 Some applications have high security

• 00:01:08 high security requirements, not critical systems, and

• 00:01:11 requirements, not critical systems, and special software engineering

• 00:01:13 special software engineering techniques can be used to achieve this. So, in which

• 00:01:16 techniques can be used to achieve this. So, in which critical systems is this necessary? Why is this necessary in

• 00:01:18 critical systems is this necessary? Why is this necessary in medical systems? Because if an error occurs,

• 00:01:21 medical systems? Because if an error occurs, human life and health are at stake.

• 00:01:24 human life and health are at stake. Telecommunications and power systems. If

• 00:01:27 Telecommunications and power systems. If communication is interrupted or power

• 00:01:29 communication is interrupted or power is cut off, it can cause very serious problems.

• 00:01:33 is cut off, it can cause very serious problems. Even electricity, for example,

• 00:01:35 Even electricity, for example,

• 00:01:41 can cause fires and other things. If space systems fail and a critical error occurs, it

• 00:01:42 systems fail and a critical error occurs, it

• 00:01:50 will be quite difficult to fix it. Because it's quite difficult to intervene in space, as you know,

• 00:01:53 quite difficult to intervene in space, as you know, sometimes they do spacewalks or

• 00:01:56 sometimes they do spacewalks or there's no battery station. And an autonomous system, you

• 00:01:59 there's no battery station. And an autonomous system, you

• 00:02:03 will never have the opportunity for physical intervention. In this way,

• 00:02:08 this way, let's see in which cases the reliability success is

• 00:02:12 let's see in which cases the reliability success is successful. If we pay the same error prevention, it is

• 00:02:15 successful. If we pay the same error prevention, it is a success you can trust.

• 00:02:17 a success you can trust. This system has been developed to avoid human error

• 00:02:20 This system has been developed to avoid human error and thus

• 00:02:22 and thus minimize system failures.

• 00:02:25 minimize system failures. This way, errors occur

• 00:02:27 This way, errors occur

• 00:02:34 during the development process. It is designed to detect and repair system failures before they are delivered to the customer. In other words,

• 00:02:42 we will organize the development process in such a way that we will detect

• 00:02:43 we will detect errors and correct them before they are delivered to the customer.

• 00:02:46 errors and correct them before they are delivered to the customer. Another element is fault detection,

• 00:02:49 Another element is fault detection, verification, and validation. Here, what is

• 00:02:54 verification, and validation. Here, what is Validation? Byndo application.

• 00:03:00 Validation? Byndo application. And what should we say here before, what we said is

• 00:03:16 verification and validation. Now, verification and validation techniques are used to detect and eliminate

• 00:03:19 verification and validation techniques are used to detect and eliminate errors in a system before they are implemented. This is

• 00:03:21 errors in a system before they are implemented. This is

• 00:03:30 a fault tolerance system. This system is always

• 00:03:32 always

• 00:03:39 designed so that errors in the software from the slime will not cause system failures. So, for example, you add a global error detection system to the

• 00:03:41 error detection system to the system.

• 00:03:44 system. Even if a failure occurs, this

• 00:03:46 Even if a failure occurs, this minimizes grease. This is both fishy. In other words, the system

• 00:03:52 minimizes grease. This is both fishy. In other words, the system

• 00:03:58 uses LO without causing a message crash, and the system continues to work, restarts automatically.

• 00:03:59 restarts automatically. Think about it now. The

• 00:04:04 Think about it now. The increasing costs of troubleshooting are increasing.

• 00:04:07 increasing costs of troubleshooting are increasing. What does "failure" mean? Recep Rezil, the

• 00:04:11 What does "failure" mean? Recep Rezil, the rezil year error, that is,

• 00:04:17 rezil year error, that is, you detect the main failures, but

• 00:04:18 you detect the main failures, but if you want to detect such rare failures,

• 00:04:21 if you want to detect such rare failures, you can. What happens to us? Friends, the

• 00:04:25 you can. What happens to us? Friends, the cost of this increases. That's why,

• 00:04:29 cost of this increases. That's why, after a certain point, the cost increases very quickly and

• 00:04:30 after a certain point, the cost increases very quickly and very much, and from there, you no

• 00:04:34 very much, and from there, you no longer seek to fix subsequent errors.

• 00:04:36 longer seek to fix subsequent errors.

• 00:04:46 This means that if this error occurs, you take the risk of the consequences of the error. This means that we can add what the cost is,

• 00:04:51 we can add what the cost is, and even systems that are natural

• 00:04:55 and even systems that are natural critical systems are regulated systems,

• 00:04:58 critical systems are regulated systems, which means that their use must be approved by

• 00:05:01 which means that their use must be approved by an external regulator before entering service.

• 00:05:03 an external regulator before entering service.

• 00:05:08 For example, there are nuclear systems,

• 00:05:10 systems, but air traffic control systems, medical

• 00:05:12 but air traffic control systems, medical devices, 300

• 00:05:15 devices, 300 systems that directly affect human health. So,

• 00:05:19 systems that directly affect human health. So, these and systems that are

• 00:05:23 these and systems that are approved by a regulator

• 00:05:24 approved by a regulator

• 00:05:35 must be approved by the regulator. Therefore, we must establish a wing to claim that a critical system development system is

• 00:05:37 development system is reliable, safe, and secure.

• 00:05:39 reliable, safe, and secure.

• 00:05:50 That is, the regulator and the regulator. More precisely, the commission must be convinced that the system

• 00:05:54 More precisely, the commission must be convinced that the system is safe, reliable, and secure.

• 00:05:56 is safe, reliable, and secure. That's why the system should be

• 00:05:58 That's why the system should be developed in this way.

• 00:06:02 developed in this way. If it can be included, a nice reporting of your voices

• 00:06:05 If it can be included, a nice reporting of your voices and the results

• 00:06:08 and the results you add here. This is famous.

• 00:06:10 you add here. This is famous. This trust, reliability, success,

• 00:06:13 This trust, reliability, success, what can we say? Methods that increase mother.

• 00:06:17 what can we say? Methods that increase mother. Isn't it my fault? It already has rules like fault detection and error detection.

• 00:06:20 Isn't it my fault? It already has rules like fault detection and error detection. Because

• 00:06:23 Because it's a system that's built here,

• 00:06:25 it's a system that's built here, you have to follow those rules. Diversity and

• 00:06:29 you have to follow those rules. Diversity and redundancy are critical. Let's combine them and

• 00:06:34 redundancy are critical. Let's combine them and have multiple versions ready.

• 00:06:36 have multiple versions ready. So, if one fails, it

• 00:06:38 So, if one fails, it can be used somewhere.

• 00:06:40 can be used somewhere. We've said before, for example, there are multiple systems in airplanes. There are

• 00:06:43 We've said before, for example, there are multiple systems in airplanes. There are two main systems,

• 00:06:46 two main systems, then subsystems, emergency

• 00:06:48 then subsystems, emergency systems, and these are developed independently of each other.

• 00:06:50 systems, and these are developed independently of each other.

• 00:06:55 If one doesn't work, the other one takes over. For example, if an airplane's engine fails, the

• 00:06:58 For example, if an airplane's engine fails, the blades

• 00:07:01 blades are designed to land with the other engine. These are all redundancy,

• 00:07:06 are designed to land with the other engine. These are all redundancy, diversity, and diversity. These provide the same functionality in different

• 00:07:08 diversity, and diversity. These provide the same functionality in different ways. So, they

• 00:07:10 ways. So, they don't fail at the same time. I

• 00:07:15 don't fail at the same time. I mentioned this. For example, the airplane's main system

• 00:07:18 mentioned this. For example, the airplane's main system was developed in two copies, but the second copy is developed

• 00:07:20 was developed in two copies, but the second copy is developed by a different team using a different

• 00:07:22 by a different team using a different programming language and different algorithms.

• 00:07:25 programming language and different algorithms. However, adding diversity and

• 00:07:28 However, adding diversity and redundancy is complex, and this can

• 00:07:31 redundancy is complex, and this can increase the risk. Of

• 00:07:33 increase the risk. Of course, what are the different things that come with this?

• 00:07:37 course, what are the different things that come with this? There are risks, and there may be dangers. Some

• 00:07:41 There are risks, and there may be dangers. Some engineers advocate for simplicity and comprehensive

• 00:07:43 engineers advocate for simplicity and comprehensive validation. Verification, or rather, validation, is

• 00:07:45 validation. Verification, or rather, validation, is

• 00:07:54 a more effective way to achieve software security. I need to do both. This means both

• 00:07:56 need to do both. This means both redundancy, diversity, and

• 00:07:59 redundancy, diversity, and verification. For Validation, diversity, redundancy,

• 00:08:02 verification. For Validation, diversity, redundancy, and redundancy are examples of diversity and redundancy. For

• 00:08:05 and redundancy are examples of diversity and redundancy. For example, in critical locations where it can be used,

• 00:08:07 example, in critical locations where it can be used, for example, in e-commerce systems.

• 00:08:09 for example, in e-commerce systems. This can also be used. The system must

• 00:08:13 also be used. The system must be constantly online. Why? Because

• 00:08:15 be constantly online. Why? Because if it's not online, it can have a significant economic cost.

• 00:08:18 if it's not online, it can have a significant economic cost. Companies normally

• 00:08:22 Companies normally maintain backup servers, and if a failure

• 00:08:24 maintain backup servers, and if a failure occurs, they automatically

• 00:08:26 occurs, they automatically switch to them. This also applies to hospital systems.

• 00:08:29 switch to them. This also applies to hospital systems. Why? Because, for example,

• 00:08:33 Why? Because, for example, during busy hours. If the hospital management

• 00:08:35 during busy hours. If the hospital management system crashes, it needs to have a server

• 00:08:38 system crashes, it needs to have a server

• 00:08:41 to process patient transactions.

• 00:08:47 Backups can be made immediately, and money can be collected periodically. However, this

• 00:08:49 be collected periodically. However, this needs to be maintained instantly.

• 00:08:52 needs to be maintained instantly. So, what needs to be a live server? It

• 00:08:54 So, what needs to be a live server? It

• 00:08:59 needs to be activated immediately. This is what redundancy diversity is. Different results can be applied using

• 00:09:02 redundancy diversity is. Different results can be applied using different operating

• 00:09:04 different operating systems, such as Windows and Linux, to provide some protection against external attacks.

• 00:09:06 systems, such as Windows and Linux, to provide some protection against external attacks.

• 00:09:10 This is for a different critical system.

• 00:09:13 for a different critical system. For example, if you're using a system vulnerable to attack,

• 00:09:17 For example, if you're using a system vulnerable to attack, you might have two servers.

• 00:09:20 you might have two servers. But one is a different operating system.

• 00:09:22 But one is a different operating system. So, it

• 00:09:26 So, it

• 00:09:31 would be more reasonable to use both redundancy and diversity simultaneously. Think about it this way.

• 00:09:37 Think about it this way. Process

• 00:09:39 Process activities like time variations and redundancy verification should be

• 00:09:41 activities like time variations and redundancy verification should be tied to a single approach, such as testing to verify the system.

• 00:09:43 tied to a single approach, such as testing to verify the system. Instead, multiple different

• 00:09:46 Instead, multiple different process activities complement each other, and

• 00:09:49 process activities complement each other, and I'll write about process errors that can lead to errors.

• 00:09:51 I'll write about process errors that can lead to errors. To prevent cross-checking,

• 00:09:53 To prevent cross-checking, tomorrow we also allow it here, for example, in

• 00:09:57 tomorrow we also allow it here, for example, in Artificial Intelligence, we are changing it to Cross Validation.

• 00:09:59 Artificial Intelligence, we are changing it to Cross Validation. What is it used for? It's

• 00:10:01 What is it used for? It's not all just for Artificial Intelligence. This

• 00:10:04 not all just for Artificial Intelligence. This song isn't essential for AI. If

• 00:10:08 song isn't essential for AI. If there's no AI,

• 00:10:10 there's no AI, what does it do in this method anyway? Friends, we develop an algorithm

• 00:10:12 what does it do in this method anyway? Friends, we develop an algorithm on different data sets,

• 00:10:15 on different data sets, and

• 00:10:18 and this is averaged. In other words,

• 00:10:21 this is averaged. In other words, a cross-validation is performed. In this way,

• 00:10:26 a cross-validation is performed. In this way,

• 00:10:38 you test whether the AI ​​you developed has memorized or whether you know what you can teach a general AI. The goal in Artificial Intelligence is to achieve generalized intelligence.

• 00:10:41 goal in Artificial Intelligence is to achieve generalized intelligence.

• 00:10:51 We can give a 100% accurate algorithm model on what is trained and on the training set, or rather, on what it is. But when it comes to

• 00:10:53 on the training set, or rather, on what it is. But when it comes to data it hasn't seen before, it

• 00:10:56 data it hasn't seen before, it

• 00:11:02 will fail because it memorizes a pill. But if it had done general learning,

• 00:11:03 But if it had done general learning, applying it would have a much higher probability.

• 00:11:06 applying it would have a much higher probability. That's how

• 00:11:10 That's how I allow software to ensure errors.

• 00:11:13 I allow software to ensure errors.

• 00:11:17 Why is it repeatable? And as long as the third is written,

• 00:11:20 repeatable? And as long as the third is written, what is a living system that

• 00:11:23 what is a living system that needs to constantly evolve and progress?

• 00:11:25 needs to constantly evolve and progress? Therefore, the process will

• 00:11:27 Therefore, the process will need to be repeated. A well-defined,

• 00:11:30 need to be repeated. A well-defined, repeatable process is

• 00:11:32 repeatable process is a process that is not entirely dependent on individual skills.

• 00:11:34 a process that is not entirely dependent on individual skills. Rather, it

• 00:11:37 Rather, it can be animated by different people. Because a

• 00:11:40 can be animated by different people. Because a person can have a well-defined, repeatable process. Regulators use information about the process to verify whether good

• 00:11:43 person can have a well-defined, repeatable process. Regulators use information about the process to verify whether good software engineering practices

• 00:11:45 software engineering practices are being used.

• 00:11:47 are being used.

• 00:11:53 Of course, the regulator won't be constantly monitoring it. What will they do?

• 00:11:55 will they do? I'll look at the processes you go through. It's

• 00:11:58 I'll look at the processes you go through. It's

• 00:12:06 clear that significant effort must be dedicated to verification and validation of the process activities. So, don't give it away. Validation is the

• 00:12:10 away. Validation is the most reliable process. The characteristics of the processes, the

• 00:12:15 most reliable process. The characteristics of the processes, the process characteristics, and the

• 00:12:17 process characteristics, and the description. There must be a sampling. It can

• 00:12:20 description. There must be a sampling. It can be documented. Why? Because the

• 00:12:22 be documented. Why? Because the regulatory system will review the

• 00:12:24 regulatory system will review the documentation.

• 00:12:27 documentation. Therefore, the methods you follow must

• 00:12:29 Therefore, the methods you follow must be documentable. The process must have a defined process model that outlines

• 00:12:33 be documentable. The process must have a defined process model that outlines the activities in the process and the documents

• 00:12:35 the activities in the process and the documents to be produced during these activities.

• 00:12:36 to be produced during these activities.

• 00:12:48 A comprehensive set of standards must be developed for standardized software production and documentation.

• 00:12:49 set of standards must be developed for standardized software production and documentation. This is an auditable

• 00:12:57 This is an auditable process. The process must be understandable to people other than participants,

• 00:12:59 process. The process must be understandable to people other than participants, allowing for verification of compliance and making

• 00:13:01 allowing for verification of compliance and making recommendations for process improvement.

• 00:13:04 recommendations for process improvement.

• 00:13:15 Various auditable processes include redundancy and various verification and validation

• 00:13:16 and various verification and validation activities. What does this

• 00:13:20 activities. What does this increase? Is it a

• 00:13:23 increase? Is it a robust process that

• 00:13:27 robust process that

• 00:13:31 can recover from the failures of individual processes? So, an individual's error should not undermine your entire

• 00:13:35 So, an individual's error should not undermine your entire process. It should not destroy it.

• 00:13:38 process. It should not destroy it. Such a process is impossible. Error 11

• 00:13:41 Such a process is impossible. Error 11

• 00:13:47 Think of these verification activities as requirements

• 00:13:52 these verification activities as requirements reviews.

• 00:13:54 reviews. But what is this verification activity?

• 00:13:56 But what is this verification activity? Requirements management, formal specifications,

• 00:14:00 Requirements management, formal specifications, system models, design and code

• 00:14:03 system models, design and code review, static analysis, test planning

• 00:14:07 review, static analysis, test planning and management. Change management, discussed in Lesson 25,

• 00:14:09 and management. Change management, discussed in Lesson 25, is not important, but

• 00:14:11 is not important, but Lesson 25 is not the subject of this period. Rock fault

• 00:14:17 Lesson 25 is not the subject of this period. Rock fault tolerance. Software

• 00:14:21 tolerance. Software systems must be fault-tolerant in critical situations.

• 00:14:23 systems must be fault-tolerant in critical situations. Because errors can also be the same,

• 00:14:26 Because errors can also be the same, what happens in the Web-based games I developed? What

• 00:14:30 what happens in the Web-based games I developed? What happens in me and polishing the server? If the

• 00:14:34 happens in me and polishing the server? If the server software encounters an error at the global level,

• 00:14:37 server software encounters an error at the global level, I and the

• 00:14:40 I and the columns of the application are running differently.

• 00:14:42 columns of the application are running differently.

• 00:14:47 What happens in this process? The application restarted, and since

• 00:14:50 The application restarted, and since it runs in these, multiple

• 00:14:52 it runs in these, multiple users can use it. I can run it. RAM memory does

• 00:14:54 users can use it. I can run it. RAM memory does

• 00:14:57 this, but there is only one shared database, so what happens if there is an error in the database?

• 00:14:59 so what happens if there is an error in the database? Vestel needs to stay.

• 00:15:01 Vestel needs to stay.

• 00:15:07 Of course, I'm not actually providing the entire ceiling for the error, but I

• 00:15:14 can increase fault tolerance by running columns in other static parts of the database. So, when writing, look at a common. More

• 00:15:22 So, when writing, look at a common. More precisely, a technology is needed, even if it's the latest,

• 00:15:24 precisely, a technology is needed, even if it's the latest, static data will be read without memory.

• 00:15:27 static data will be read without memory. Bre fault tolerance,

• 00:15:32 Bre fault tolerance, highest availability.

• 00:15:34 highest availability.

• 00:15:42 It's necessary in places where there are high requirements or the costs of system failure are very high. Fault tolerance

• 00:15:48 means that the system can continue to function despite software failure. So, how do you do this? If you

• 00:15:53 have a Windows application and an error occurs, you

• 00:15:57 can't catch it. If you haven't anticipated it before, your application might give an error and

• 00:16:01 you haven't anticipated it before, your application might give an error and stop working.

• 00:16:03 stop working.

• 00:16:10 You've probably seen errors like "stop working." So, what can you do

• 00:16:12 can you do to prevent this from stopping? For example, by adding a global error-

• 00:16:15 to prevent this from stopping? For example, by adding a global error- catching method to the software,

• 00:16:19 catching method to the software,

• 00:16:25 you can increase the error-recovery function. In this case, you can ensure that the application

• 00:16:29 continues to function in a different way.

• 00:16:35 This is the mother of this. If you start the methods and functions in the programming language

• 00:16:44 as subtasks, since errors won't occur in the main task you're writing, the

• 00:16:46 errors won't occur in the main task you're writing, the errors in the subtasks

• 00:16:49 errors in the subtasks won't prevent the main software from running. You can

• 00:16:53 won't prevent the main software from running. You can restart that subtask when an error occurs.

• 00:16:55 restart that subtask when an error occurs. This also

• 00:16:59 This also raises the bar. Even if the system is proven to comply with the specifications, it

• 00:17:03 raises the bar. Even if the system is proven to comply with the specifications, it

• 00:17:16 must be fault-tolerant, as specification errors may occur or validation may be incorrect. A

• 00:17:21 reliable system architecture is used in situations where error detection is necessary.

• 00:17:23 in situations where error detection is necessary. These architectures are generally

• 00:17:27 These architectures are generally redundant and have different types. What are

• 00:17:29 redundant and have different types. What are cloning and cloning, plus the difference?

• 00:17:35 cloning and cloning, plus the difference? Do I offer this diversity?

• 00:17:38 Do I offer this diversity? Examples of situations where these situations are used are

• 00:17:40 Examples of situations where these situations are used are

• 00:17:46 flight control systems, where a system failure can threaten the safety of travel. Reactor systems, where a control system failure

• 00:17:48 control system failure

• 00:17:54 can lead to a chemical or nuclear emergency, require 24/7 availability.

• 00:17:56 require 24/7 availability. Telecommunication systems, protection

• 00:18:02 Telecommunication systems, protection systems, a system specific to

• 00:18:05 systems, a system specific to another control system that can provide emergency response when a malfunction occurs. A

• 00:18:07 another control system that can provide emergency response when a malfunction occurs. A

• 00:18:15 system that stops the train if it passes a red light, but is independent of the temperature and pressure. A system that shuts down the train at night is very loud.

• 00:18:17 pressure. A system that shuts down the train at night is very loud.

• 00:18:21 These licenses also control the system. Independent protection systems monitor the

• 00:18:25 Independent protection systems monitor the controlled system and the environment independently.

• 00:18:27 controlled system and the environment independently. It is important to be independent because

• 00:18:30 It is important to be independent because if it is dependent, it may not be activated because it will be

• 00:18:32 if it is dependent, it may not be activated because it will be activated when a malfunction occurs in the system to which it is dependent. If

• 00:18:34 activated when a malfunction occurs in the system to which it is dependent. If

• 00:18:40 a problem is detected, it takes emergency action to shut down the system or to avoid a disaster. It

• 00:18:43 shut down the system or to avoid a disaster. It

• 00:18:48 gives commands. What does the protection system give this command? What is the protection system architecture?

• 00:18:54 this command? What is the protection system architecture? Friends, look at what is there.

• 00:18:56 Friends, look at what is there. There is a system, so the system works in common.

• 00:19:00 There is a system, so the system works in common. This is the only one. Now, you are SOS, meaning it

• 00:19:02 This is the only one. Now, you are SOS, meaning it will provide protection. You are the questions themselves.

• 00:19:05 will provide protection. You are the questions themselves. What is under that? There is a

• 00:19:07 What is under that? There is a protexion system based on sensor data.

• 00:19:09 protexion system based on sensor data. What is a protexion system? It is an activity, meaning the

• 00:19:13 What is a protexion system? It is an activity, meaning the physical elements that will perform that protection process.

• 00:19:15 physical elements that will perform that protection process. For example, if

• 00:19:17 For example, if we sewed the temperature, it could be a simple electronic circuit system that turns off the colors if it is very high.

• 00:19:21 we sewed the temperature, it could be a simple electronic circuit system that turns off the colors if it is very high. What is the system that starts?

• 00:19:23 What is the system that starts?

• 00:19:30 For example, we gave an example of a dam. There might be a system that opens the dam gates.

• 00:19:32 might be a system that opens the dam gates. Think about it all. Isn't there this control addition at the bottom?

• 00:19:37 Think about it all. Isn't there this control addition at the bottom?

• 00:19:44 What are the controlled tools and equipment? They're connected to sensors and

• 00:19:48 They're connected to sensors and sending them. More precisely, on the right,

• 00:19:50 sending them. More precisely, on the right, there are sensors. There's already

• 00:19:53 there are sensors. There's already a control system from the questions. It

• 00:19:54 a control system from the questions. It continues in the same way.

• 00:19:59 continues in the same way. There's a protection system like this. The sensor on the right is the

• 00:20:02 There's a protection system like this. The sensor on the right is the protected system. The sensor on the left is the system that protects the sensors.

• 00:20:06 protected system. The sensor on the left is the system that protects the sensors. The sensors are

• 00:20:16 a protection system. The functionality of the protection systems is unnecessary because they include

• 00:20:19 systems is unnecessary because they include

• 00:20:24 monitoring and control capabilities that replicate those in the control software.

• 00:20:36 Look carefully here. I think the meaning of the unnecessary base is that it's not functional right now. Let's

• 00:20:38 I think the meaning of the unnecessary base is that it's not functional right now. Let's see, let's say it's

• 00:20:55 see, let's say it's not unnecessary. What is it redundant? Why? Because the

• 00:21:02 not unnecessary. What is it redundant? Why? Because the protection system already

• 00:21:05 protection system already

• 00:21:09 includes the features in our current system. It doesn't provide us with any extra benefit other than protecting how much protection is protected. In other words, it

• 00:21:13 than protecting how much protection is protected. In other words, it

• 00:21:19 's actually a non-functional feature. Protection systems should be diverse and control systems should use at least different

• 00:21:21 should be diverse and control systems should use at least different technologies.

• 00:21:26 technologies. Since they are more than the control system,

• 00:21:27 Since they are more than the control system,

• 00:21:31 more effort can be spent on verification. This hole assurance is a goal to ensure

• 00:21:35 can be spent on verification. This hole assurance is a goal to ensure a low probability of failure on demand for the protection system.

• 00:21:37 a low probability of failure on demand for the protection system. Here,

• 00:21:47 Here, on demand. What happened, friends?

• 00:21:51 on demand. What happened, friends? When an unexpected situation occurs, it just

• 00:21:53 When an unexpected situation occurs, it just kicks in. It doesn't work continuously. It

• 00:21:56 kicks in. It doesn't work continuously. It doesn't want anything that isn't added.

• 00:21:59 doesn't want anything that isn't added.

• 00:22:07 Subscribe to self-monitoring architectures. This system monitors its own processes

• 00:22:16 This system monitors its own processes and detects these inconsistencies.

• 00:22:19 and detects these inconsistencies. Because multi-channel

• 00:22:21 Because multi-channel architectures, the same calculations are made on each channel

• 00:22:25 architectures, the same calculations are made on each channel and the results are compared. If

• 00:22:28 and the results are compared. If the results are the same, also if the subway

• 00:22:30 the results are the same, also if the subway system worked correctly, see

• 00:22:32 system worked correctly, see what's going on here.

• 00:22:39 what's going on here. For example, there are two sensors

• 00:22:42 For example, there are two sensors coming from two of them. Think of it as three sensors. If the results

• 00:22:45 coming from two of them. Think of it as three sensors. If the results from two of the three are the same, it continues

• 00:22:47 from two of the three are the same, it continues or decides on it. If

• 00:22:50 or decides on it. If the results are different, a failure is

• 00:22:52 the results are different, a failure is assumed and a failure exception

• 00:22:55 assumed and a failure exception occurs. This is a self-

• 00:22:58 occurs. This is a self- monitoring system from us. Hacer, self-monitoring

• 00:23:01 monitoring system from us. Hacer, self-monitoring Architect, look like this: there's an input, the

• 00:23:03 Architect, look like this: there's an input, the input is divided into two. Channel 1, Channel 2,

• 00:23:06 input is divided into two. Channel 1, Channel 2, Channel 1,

• 00:23:07 Channel 1, the results from this country

• 00:23:09 the results from this country are compared. If they're the same,

• 00:23:12 are compared. If they're the same, I said okay, the system is working correctly. If

• 00:23:14 I said okay, the system is working correctly. If they're different, there's an error in the system.

• 00:23:16 they're different, there's an error in the system. These self-monitoring

• 00:23:21 These self-monitoring systems require different hardware on each channel to prevent common mode hardware failures from

• 00:23:25 systems require different hardware on each channel to prevent common mode hardware failures from producing the same results.

• 00:23:26 producing the same results.

• 00:23:31 This is very important. There will be no picket clones on channels because

• 00:23:35 There will be no picket clones on channels because an error or vulnerability in one will

• 00:23:38 an error or vulnerability in one will also be transmitted to the other. If it's used, it's assumed there's a failure.

• 00:23:41 also be transmitted to the other. If it's used, it's assumed there's a failure.

• 00:23:46 There should be variety in their writings. Otherwise, the same software error will affect the manual channel. That is, both the

• 00:23:48 error will affect the manual channel. That is, both the hardware and the software must be written differently so that it

• 00:23:50 hardware and the software must be written differently so that it works. This way, both

• 00:23:55 works. This way, both software and hardware are

• 00:23:58 software and hardware are prevented from failing simultaneously with the same error. Because there

• 00:24:01 prevented from failing simultaneously with the same error. Because there is diversity, if high availability is

• 00:24:04 is diversity, if high availability is required, you can use a few

• 00:24:07 required, you can use a few self-monitoring systems in parallel.

• 00:24:09 self-monitoring systems in parallel. This is the approach used

• 00:24:13 This is the approach used for flight control systems in the Erva Hot family.

• 00:24:15 for flight control systems in the Erva Hot family. Because there is

• 00:24:19 Because there is both redundancy and diversity.

• 00:24:21 both redundancy and diversity. Look, we will see now. If the main flight

• 00:24:25 Look, we will see now. If the main flight control system Cave has an input,

• 00:24:28 control system Cave has an input, friends, the input

• 00:24:31 friends, the input is transmitted to all systems. The same input is compared to the

• 00:24:34 is transmitted to all systems. The same input is compared to the main flight

• 00:24:37 main flight control system.

• 00:24:39 control system.

• 00:24:50 There are two more channels. These are backup systems. They can also

• 00:24:54 These are backup systems. They can also process the input and produce the result if necessary.

• 00:24:59 process the input and produce the result if necessary. Then there is a second-hand

• 00:25:03 Then there is a second-hand flight system. The second Rüstem also

• 00:25:06 flight system. The second Rüstem also

• 00:25:13 has that self-monitoring system. Now we will see in more detail. If the

• 00:25:17 Now we will see in more detail. If the main architecture discussion. If you look, TCS

• 00:25:20 main architecture discussion. If you look, TCS

• 00:25:28 has five separate computers that can run any control software. 5 of them, and only one of them actually performs its flight. Various

• 00:25:30 actually performs its flight. Various types are

• 00:25:31 types are used extensively. What are

• 00:25:34 used extensively. What are these first systems, which have

• 00:25:39 these first systems, which have a different processor than their systems?

• 00:25:41 a different processor than their systems?

• 00:25:48 The software in each channel is developed by different teams in different programming languages. The

• 00:26:13 different programming languages

• 00:26:19 used in the first and second systems are different. Thus, the probability of

• 00:26:30 two teams making the same algorithmic design error is greatly

• 00:26:33 error is greatly reduced.

• 00:26:34 reduced. Subscribe.

• 00:26:36 Subscribe. Key points in this section are as

• 00:26:38 Key points in this section are as follows: reliability in a program is

• 00:26:41 follows: reliability in a program is preventing errors from occurring. By

• 00:26:44 preventing errors from occurring. By detecting and eliminating errors before system deployment, the system provides

• 00:26:47 detecting and eliminating errors before system deployment, the system provides comfortable land

• 00:26:50 comfortable land opportunities. We can

• 00:26:54 opportunities. We can

• 00:27:02 also say that reliability can be achieved by including hardware, software processes, and software

• 00:27:05 hardware, software processes, and software systems. Using redundancy and diversity

• 00:27:07 systems. Using redundancy and diversity

• 00:27:13 is essential for developing systems. It is crucial to minimize system failures by using a

• 00:27:15 minimize system failures by using a well-defined, repeatable

• 00:27:18 well-defined, repeatable process.

• 00:27:21 process. Secure system architectures

• 00:27:24 Secure system architectures are system architectures designed for error management.

• 00:27:27 are system architectures designed for error management. See fault tolerance. This error is

• 00:27:32 See fault tolerance. This error is between 80 architectures and styles. Protection

• 00:27:35 between 80 architectures and styles. Protection systems.

• 00:27:36 systems. O self-monitoring exchanges and the most

• 00:27:38 O self-monitoring exchanges and the most difficult version of programming. Find the most difficult programming.

• 00:27:40 difficult version of programming. Find the most difficult programming.

• 00:27:48 Yes, we continue in section 2. The most reliable program. Multiple versions of a software system

• 00:27:51 program. Multiple versions of a software system perform the same calculations.

• 00:27:54 perform the same calculations. There should be an odd number of computers within the work.

• 00:27:56 There should be an odd number of computers within the work. Typically, 3. So, from an odd number,

• 00:27:59 Typically, 3. So, from an odd number, results like 35 to 59

• 00:28:03 results like 35 to 59 are compared using a voting system, and the majority

• 00:28:05 are compared using a voting system, and the majority result is taken as the correct result. The

• 00:28:07 result is taken as the correct result. The voting approach we gave is the

• 00:28:10 voting approach we gave is the triple modular concept used in hardware systems. It

• 00:28:12 triple modular concept used in hardware systems. It is now produced according to the concept of Banan

• 00:28:17 is now produced according to the concept of Banan fault tolerance. Models are

• 00:28:19 fault tolerance. Models are redundant. TMR depends on trimo. Dur

• 00:28:26 redundant. TMR depends on trimo. Dur Rivers can be. Tiribolulu. We saw it

• 00:28:32 Rivers can be. Tiribolulu. We saw it in English. Let's look at them right away.

• 00:28:36 in English. Let's look at them right away. xDrive. Modern residence. Yes, redundancy is the

• 00:28:39 xDrive. Modern residence. Yes, redundancy is the same. Comparing the input and output fields, there

• 00:28:42 same. Comparing the input and output fields, there

• 00:28:46 are 33 components. If an output is different, it's ignored and component

• 00:28:48 is different, it's ignored and component failure is considered present. Instead of design errors,

• 00:28:51 failure is considered present. Instead of design errors, you'll know that

• 00:28:53 you'll know that most errors resulting from failures, or simultaneous

• 00:28:55 most errors resulting from failures, or simultaneous component failures, are

• 00:28:57 component failures, are based on the fact that the majority of these errors are considered. When this happens,

• 00:29:02 based on the fact that the majority of these errors are considered. When this happens, of course, the output of the majority is accepted.

• 00:29:06 of course, the output of the majority is accepted. Here, you see the multitude of programs, there

• 00:29:09 Here, you see the multitude of programs, there are different versions. You see, the

• 00:29:10 are different versions. You see, the majority of these are accepted as outputs.

• 00:29:13 majority of these are accepted as outputs. We'll add them, sometimes it hurts.

• 00:29:17 We'll add them, sometimes it hurts. I have your program. It was designed and implemented

• 00:29:20 I have your program. It was designed and implemented by teams in the park for different system versions.

• 00:29:22 by teams in the park for different system versions. Look, there

• 00:29:24 Look, there 's an important nuance here. It's assumed that the

• 00:29:28 's an important nuance here. It's assumed that the same errors are unlikely to occur by different teams.

• 00:29:31 same errors are unlikely to occur by different teams. Naturally, the

• 00:29:33 Naturally, the algorithms used should be different,

• 00:29:36 algorithms used should be different, but they might not be different. Here,

• 00:29:41 but they might not be different. Here, they also use different programming languages.

• 00:29:43 they also use different programming languages. This

• 00:29:46 This reduces the probability of algorithmic errors. There's some empirical evidence that teams often

• 00:29:50 reduces the probability of algorithmic errors. There's some empirical evidence that teams often misinterpret specifications in the same way

• 00:29:52 misinterpret specifications in the same way and choose the same algorithms in their systems.

• 00:29:54 and choose the same algorithms in their systems.

• 00:30:00 Yes, this can happen, but the

• 00:30:03 this can happen, but the probability is very low.

• 00:30:05 probability is very low. When this happens,

• 00:30:09 When this happens, there are various methods to reduce this. We

• 00:30:11 there are various methods to reduce this. We 'll explain them within the software types. We'll discuss

• 00:30:14 'll explain them within the software types. We'll discuss approaches to software errors within the

• 00:30:17 approaches to software errors within the same software specification. So,

• 00:30:19 same software specification. So, you say, for example, when 05 pressure

• 00:30:23 you say, for example, when 05 pressure is measured, here's a head tree. So, I

• 00:30:31 is measured, here's a head tree. So, I 'm giving a random example here, or when

• 00:30:32 'm giving a random example here, or when blood sugar is measured, there

• 00:30:35 blood sugar is measured, there 's this much dose. What is this? The software

• 00:30:39 's this much dose. What is this? The software specification defines the task. But the algorithms and programming language that determine

• 00:30:42 specification defines the task. But the algorithms and programming language that determine how this task will be performed

• 00:30:43 how this task will be performed

• 00:30:52 depend entirely on the developer. This means that the program code can be written in many different ways,

• 00:30:55 can be written in many different ways,

• 00:31:01 depending on the hardware and programming language to be used. This specification here

• 00:31:06 will intentionally fail in different ways. It depends on the software types, it is independent of the applications, and

• 00:31:10 is independent of the applications, and

• 00:31:15 it is assumed that it does not contain common errors. Strategies for achieving diversity include different programming

• 00:31:17 Strategies for achieving diversity include different programming languages, different design methods and

• 00:31:19 languages, different design methods and tools, and clear specification of different algorithms.

• 00:31:22 tools, and clear specification of different algorithms.

• 00:31:28 After doing all of this, the difference is that different applications will make the same mistake,

• 00:31:31 applications will make the same mistake, and it will take very little time.

• 00:31:35 and it will take very little time. Problems related to design types are

• 00:31:38 Problems related to design types are not different in terms of team culture.

• 00:31:40 not different in terms of team culture. Therefore, what is the action of addressing problems in the same way?

• 00:31:42 Therefore, what is the action of addressing problems in the same way?

• 00:31:46 Yes, this also applies to how teams are selected when forming teams,

• 00:31:54 team members, and Bekir's subordinates will also make a significant difference. Everyone

• 00:31:56 subordinates will also make a significant difference. Everyone comes from the same or similar environments and

• 00:31:59 comes from the same or similar environments and makes the same mistake. What can we say?

• 00:32:04 makes the same mistake. What can we say? It can lead to a predisposition to characteristic errors.

• 00:32:07 It can lead to a predisposition to characteristic errors. Different teams make the same mistakes.

• 00:32:10 Different teams make the same mistakes. Some parts of an application make the same mistakes

• 00:32:12 Some parts of an application make the same mistakes later than others. Therefore, all

• 00:32:14 later than others. Therefore, all teams tend to make mistakes in the same place.

• 00:32:15 teams tend to make mistakes in the same place. What are characteristic attacks in terms of

• 00:32:17 What are characteristic attacks in terms of specification errors? If the

• 00:32:20 specification errors? If the specification is faulty, the

• 00:32:23 specification is faulty, the team will make mistakes. There's nothing I can do about it.

• 00:32:25 team will make mistakes. There's nothing I can do about it. If there's a mistake in the specification, it will

• 00:32:28 If there's a mistake in the specification, it will either heat up all applications. Because mold, even

• 00:32:32 either heat up all applications. Because mold, even if it's with different teams. They

• 00:32:35 if it's with different teams. They have to monitor and implement that specification. They

• 00:32:38 have to monitor and implement that specification. They have to develop it. This can also be done to

• 00:32:41 have to develop it. This can also be done to some extent using multiple features.

• 00:32:42 some extent using multiple features.

• 00:32:49 Yes, specification dependency

• 00:32:52 dependency

• 00:32:57 is open to specification errors with both approaches to software redundancy. If the specification is incorrect, the system can fail.

• 00:32:59 is incorrect, the system can fail. This is also a hardware-related

• 00:33:03 This is also a hardware-related issue. However, software features are

• 00:33:04 issue. However, software features are generally more complex than hardware features

• 00:33:06 generally more complex than hardware features and are more difficult to verify.

• 00:33:10 and are more difficult to verify. So, in hardware,

• 00:33:12 So, in hardware, errors can arise from specification dependency.

• 00:33:14 errors can arise from specification dependency. This has sometimes been addressed by developing

• 00:33:17 This has sometimes been addressed by developing separate software specifications for the same user specification. In

• 00:33:19 separate software specifications for the same user specification. In

• 00:33:24 some cases, they've even said, "Let's make the specification different, so

• 00:33:28 even said, "Let's make the specification different, so

• 00:33:35 let's try to eliminate the possibility of thinking about the same error." But of course, there's a very extreme case.

• 00:33:39 But of course, there's a very extreme case. So, how are practical and applied

• 00:33:40 So, how are practical and applied improvements made? If the principle

• 00:33:43 improvements made? If the principle is that interoperability and independence can be achieved,

• 00:33:46 and independence can be achieved, multi-version programming can lead to

• 00:33:49 multi-version programming can lead to significant improvements in reliability and usability. The

• 00:33:50 significant improvements in reliability and usability. The

• 00:33:56 improvements observed in practice are much less significant, and only the approach achieves a reliability improvement between 5 and 9 times.

• 00:33:59 significant, and only the approach achieves a reliability improvement between 5 and 9 times.

• 00:34:04 Looking at the observed improvements in practice, we can say that the observed improvements in practice are much less significant.

• 00:34:07 observed improvements in practice, we can say that the observed improvements in practice are much less significant. Because these errors

• 00:34:11 Because these errors don't happen all the time,

• 00:34:13 don't happen all the time, but when they do, they can lead to very Behçet results.

• 00:34:16 but when they do, they can lead to very Behçet results. So,

• 00:34:20 So, there's very little benefit in monitoring. We can observe what we've done,

• 00:34:21 there's very little benefit in monitoring. We can observe what we've done, but

• 00:34:25 but since the error is prevented from ever occurring, we can

• 00:34:27 since the error is prevented from ever occurring, we can actually observe how much has been gained.

• 00:34:29 actually observe how much has been gained. So, think about it this way. The

• 00:34:32 So, think about it this way. The key question is, we won't say that such improvements don't cost

• 00:34:36 key question is, we won't say that such improvements don't cost significant extra development for programming multiple versions. Of

• 00:34:38 significant extra development for programming multiple versions. Of

• 00:34:43 course, another question is that the difference and improvement

• 00:34:47 the difference and improvement will bring a very serious cost. So,

• 00:34:50 will bring a very serious cost. So, only if it's worth it and sustainable. You

• 00:34:53 only if it's worth it and sustainable. You

• 00:35:00 can make a sustainable Sancak in terms of cost and profit, otherwise we won't make that software at all. So, if the

• 00:35:02 we won't make that software at all. So, if the profit and loss ratio doesn't meet the cost,

• 00:35:07 profit and loss ratio doesn't meet the cost, reliable programming programs

• 00:35:10 reliable programming programs

• 00:35:13 help reduce the frequency of errors. These programming practices can

• 00:35:14 These programming practices can be identified. This program

• 00:35:17 be identified. This program supports the following: error

• 00:35:20 supports the following: error detection. We will do these. We've already seen these

• 00:35:23 detection. We will do these. We've already seen these in a section.

• 00:35:35 Let's look at the recommendations for the most reliable program. Why are reliable programming

• 00:35:37 Why are reliable programming guidelines? Limit the visibility of information in a program,

• 00:35:40 guidelines? Limit the visibility of information in a program,

• 00:35:47 that is, unnecessary information should not be visible.

• 00:35:51 should not be visible. Check the validity of all inputs. This is always,

• 00:35:54 Check the validity of all inputs. This is always, always, what a software

• 00:35:57 always, what a software engineer should do. You

• 00:36:02 engineer should do. You should never trust any input. You should never trust the message.

• 00:36:04 should never trust any input. You should never trust the message. String in experience. You expect a fig. Are you

• 00:36:07 String in experience. You expect a fig. Are you expecting a strip? Are you expecting that

• 00:36:08 expecting a strip? Are you expecting that city to be as you expect?

• 00:36:11 city to be as you expect?

• 00:36:20 Provide a handler for all exceptions. So, if it happens in Istanbul, that exception will be on its own. It shouldn't be left open,

• 00:36:23 exception will be on its own. It shouldn't be left open, and that exception should be managed in a way that

• 00:36:27 and that exception should be managed in a way that minimizes its use,

• 00:36:30 minimizes its use, making it error-prone. It provides restart capabilities.

• 00:36:32 making it error-prone. It provides restart capabilities. Restarting is

• 00:36:34 Restarting is crucial because its homework can cause errors, and the only

• 00:36:36 crucial because its homework can cause errors, and the only solution is restarting. And your

• 00:36:39 solution is restarting. And your software should be able to do this. Check the array boundaries. If

• 00:36:42 software should be able to do this. Check the array boundaries. If you're using an array, you

• 00:36:46 you're using an array, you

• 00:36:55 should be careful about the boundaries there. If you're manipulating elements in our indexes, you should be careful about the boundaries. When we're looking for external components, it should also be opened from time to time.

• 00:36:57 also be opened from time to time. Why? Because the external component

• 00:37:01 Why? Because the external component doesn't have control over its existence. Its response isn't in your hands. You'll

• 00:37:03 doesn't have control over its existence. Its response isn't in your hands. You'll

• 00:37:08 have to give it instructions, otherwise you might have to wait forever.

• 00:37:10 might have to wait forever.

• 00:37:17 Name all constants that represent real-world values. So, when you define a variable,

• 00:37:23 don't give it meaningless names like a, b, c, d, but what does it do? What does it do? It's

• 00:37:25 What does it do? It's something. My name is something. Then,

• 00:37:28 something. My name is something. Then,

• 00:37:34 variable naming is very important. It's very important to see information in a program.

• 00:37:36 see information in a program. Now, we

• 00:37:38 Now, we 'll explore these techniques more closely. Program

• 00:37:40 'll explore these techniques more closely. Program components should only be allowed to

• 00:37:43 components should only be allowed to access the data they need for these applications.

• 00:37:45 access the data they need for these applications. This means that

• 00:37:48 This means that some parts of the program state shouldn't be

• 00:37:50 some parts of the program state shouldn't be accidentally corrupted by these components.

• 00:37:52 accidentally corrupted by these components. We can control the data

• 00:37:55 We can control the data presentation using abstract data

• 00:37:58 presentation using abstract data types where data is specific and visible

• 00:38:01 types where data is specific and visible only. We can control the data

• 00:38:05 only. We can control the data beforehand using predefined data types like get and put. Do

• 00:38:06 beforehand using predefined data types like get and put. Do you allow access with defined operations?

• 00:38:10 you allow access with defined operations? Of course, more advanced programming is needed.

• 00:38:13 Of course, more advanced programming is needed. How to do this

• 00:38:15 How to do this can vary depending on different programming languages. All

• 00:38:18 can vary depending on different programming languages. All

• 00:38:28 programs check all inputs for validity. This is a lot of people. All programs take input from their environment and make assumptions about these inputs.

• 00:38:31 and make assumptions about these inputs. However, program

• 00:38:35 However, program features are kept secret with these assumptions.

• 00:38:38 features are kept secret with these assumptions. Otherwise, they define what to do.

• 00:38:40 Otherwise, they define what to do. Consequently, many programs

• 00:38:44 Consequently, many programs

• 00:38:49 behave unpredictably when presented with external inputs, and sometimes these are threats to system security. For

• 00:38:51 these are threats to system security. For example,

• 00:38:54 example, you will investigate this. If

• 00:38:57 you will investigate this. If

• 00:39:02 you do not verify the inputs you receive from the user, they can easily add them to our database. Oh, they even

• 00:39:05 add them to our database. Oh, they even know you completely, you are always standing still.

• 00:39:08 know you completely, you are always standing still. Here is all the information about this minced meat has

• 00:39:10 Here is all the information about this minced meat has been stolen. These are generally caused by not checking the locations and ensuring

• 00:39:14 been stolen. These are generally caused by not checking the locations and ensuring adequate control.

• 00:39:16 adequate control. Wars. As a result,

• 00:39:20 Wars. As a result,

• 00:39:23 according to the assumptions made about these inputs, you should check this before processing the process. That is, you will not

• 00:39:30 say "Input is correct." You will always think that the user is a malicious hacker.

• 00:39:33 a malicious hacker. We will check the valid

• 00:39:37 We will check the valid checks, such as the interval checks.

• 00:39:40 checks, such as the interval checks. Check the 16 known inputs of your

• 00:39:43 Check the 16 known inputs of your mother. We said, for example, a water

• 00:39:47 mother. We said, for example, a water treatment system, a chemical, for example, that

• 00:39:51 treatment system, a chemical, for example, that allows our hand to be entered, which is

• 00:39:53 allows our hand to be entered, which is harmful to human health. So,

• 00:39:56 harmful to human health. So, whether it is quoted or not, but Check You will

• 00:39:58 whether it is quoted or not, but Check You will

• 00:40:05 check that the game controls do not open the maximum size and open the Love size. For example, if our database

• 00:40:09 our database supports a maximum of 40 characters for a horse, and you

• 00:40:10 supports a maximum of 40 characters for a horse, and you allow a character, this will cause a database error. If you allow a character, this will

• 00:40:12 allow a character, this will cause a database error. If you allow a character, this will

• 00:40:17 cause a failure of five. Nurse controls check that the initiative representation

• 00:40:20 controls check that the initiative representation

• 00:40:23 does not contain characters that should not be part of it. It should not contain numbers. For example,

• 00:40:27 should not contain numbers. For example, no name in Türkiye can contain numbers.

• 00:40:29 no name in Türkiye can contain numbers. Then you will check this.

• 00:40:31 Then you will check this. Bağkurlu controls use the information about the back to

• 00:40:35 Bağkurlu controls use the information about the back to check that it is reasonable rather than an excessive value.

• 00:40:37 check that it is reasonable rather than an excessive value.

• 00:40:43 This will of course change the situation. A control that

• 00:40:47 change the situation. A control that provides a handler for all exceptions. A program

• 00:40:50 provides a handler for all exceptions. A program exception is an error or

• 00:40:52 exception is an error or some unexpected event, such as a power outage.

• 00:40:53 some unexpected event, such as a power outage. Exception handling structures

• 00:40:58 Exception handling structures

• 00:41:05 allow for the handling of such events without the need for continuous status checking to detect exceptions. This allows for the

• 00:41:09 use of normal checking to detect exceptions.

• 00:41:14 This requires many additional expressions to be added to the program. This is a significant shortcoming and potentially open to error. Bre

• 00:41:19 and potentially open to error. Bre exception handling. See what Kot Seleks's

• 00:41:21 exception handling. See what Kot Seleks's normal floof control is.

• 00:41:23 normal floof control is. This normal flow is

• 00:41:27 This normal flow is what it does when an error is detected. See what it does here, what it

• 00:41:29 what it does when an error is detected. See what it does here, what it acts as below, what this

• 00:41:31 acts as below, what this exception process is, that is, the error handling

• 00:41:35 exception process is, that is, the error handling here. Hatay processes Korkmaz: Of course,

• 00:41:39 here. Hatay processes Korkmaz: Of course, we can no longer perform normal exits. When an error

• 00:41:41 we can no longer perform normal exits. When an error occurs, our error-handling code runs,

• 00:41:44 occurs, our error-handling code runs, and what do you do next?

• 00:41:46 and what do you do next? Normal flow

• 00:41:49 Normal flow is disrupted, as is the case with the exception process. The exception process

• 00:41:54 is disrupted, as is the case with the exception process. The exception process has a strategy of requesting three exceptions. It

• 00:41:56 has a strategy of requesting three exceptions. It

• 00:42:01 notifies a calling component that an exception has occurred and provides information about the exception type, meaning it

• 00:42:05 provides information about the exception type, meaning it provides information about who called the method.

• 00:42:07 provides information about who called the method. Some alternative actions are performed to the process where the exception occurred.

• 00:42:10 Some alternative actions are performed to the process where the exception occurred. This is

• 00:42:13 This is only possible if the Istan handler

• 00:42:16 only possible if the Istan handler

• 00:42:21 has sufficient information to overcome the problem. In the case of an "excuse me," there

• 00:42:23 In the case of an "excuse me," there is a need to know which exception event

• 00:42:26 is a need to know which exception event occurred and to identify

• 00:42:28 occurred and to identify it. We need to predict it. I can

• 00:42:31 it. We need to predict it. I can write an alternative process

• 00:42:33 write an alternative process to celebrate it. For this, I can pass control to

• 00:42:36 to celebrate it. For this, I can pass control to a runtime support

• 00:42:38 a runtime support system to handle it more uniquely. Exception handling is

• 00:42:42 system to handle it more uniquely. Exception handling is a mechanism that provides some kind of ceremony for this

• 00:42:48 a mechanism that provides some kind of ceremony for this error. This minimizes the use of explicit structures.

• 00:42:50 error. This minimizes the use of explicit structures. Which structures

• 00:42:53 Which structures can be vulnerable to Kahta? Program errors are often

• 00:42:55 can be vulnerable to Kahta? Program errors are often the result of human error. This is because

• 00:42:58 the result of human error. This is because programmers lose

• 00:43:01 programmers lose track of the relationships between different parts of the system.

• 00:43:03 track of the relationships between different parts of the system. This is inherently complex,

• 00:43:09 This is inherently complex, or the ability to control errors

• 00:43:11 or the ability to control errors in programming languages ​​ like Hatay

• 00:43:13 in programming languages ​​like Hatay is exacerbated by explicit structures.

• 00:43:16 is exacerbated by explicit structures. Therefore, when programming,

• 00:43:18 Therefore, when programming,

• 00:43:23 you should try to avoid explicit structures or at least minimize their use. Error structures also go unconditionally.

• 00:43:28 Error structures also go unconditionally. So, without any conditions, we

• 00:43:32 So, without any conditions, we say, "Go directly to this line in this code." This is

• 00:43:36 say, "Go directly to this line in this code." This is not recommended at all. You should avoid this in programming.

• 00:43:39 not recommended at all. You should avoid this in programming.

• 00:43:45 Floating-point numbers are very difficult to express 100%.

• 00:43:47 very difficult to express 100%. For example, it's very difficult to click on a number like this in Word.

• 00:43:51 For example, it's very difficult to click on a number like this in Word.

• 00:43:57 Then, let me show you from there. Then, like this, like this, like this.

• 00:43:59 Then, like this, like this, like this. You can't express it exactly. That's why this

• 00:44:03 You can't express it exactly. That's why this rounding occurs.

• 00:44:05 rounding occurs. For example, if you compare this exactly, imagine this

• 00:44:07 For example, if you compare this exactly, imagine this database example is stored like this.

• 00:44:09 database example is stored like this. Now,

• 00:44:12 Now, when you encounter these two numbers, the same number doesn't come out.

• 00:44:14 when you encounter these two numbers, the same number doesn't come out. So, using floating-point numbers can

• 00:44:18 So, using floating-point numbers can lead to invalid comparisons.

• 00:44:22 lead to invalid comparisons. Comparing these two numbers

• 00:44:23 Comparing these two numbers will yield good results. But this number and

• 00:44:26 will yield good results. But this number and OK, it may be stored like this. This is

• 00:44:29 OK, it may be stored like this. This is

• 00:44:38 related to the size of the data type used with bits and bands, the precision of the data top, and pointers have signs. That is, pointers

• 00:44:41 pointers have signs. That is, pointers

• 00:44:46 can corrupt data from the sign under the wrong memory areas. A little bit can make it difficult to understand the program's operation,

• 00:44:49 little bit can make it difficult to understand the program's operation, taxing it, and changing it.

• 00:44:51 taxing it, and changing it. Dynamic memory allocation can

• 00:44:54 Dynamic memory allocation can cause memory overflows at runtime.

• 00:44:58 cause memory overflows at runtime. This can also cause errors. This can be very pleasing. Open structures,

• 00:45:03 This can also cause errors. This can be very pleasing. Open structures, parallelism, unforeseen interactions between parallel processes,

• 00:45:06 parallelism, unforeseen interactions between parallel processes,

• 00:45:10 can cause subtle errors. So, it's a very serious error. Synchronization is

• 00:45:12 So, it's a very serious error. Synchronization is required in parallel programming.

• 00:45:15 required in parallel programming.

• 00:45:22 Understanding recursion functions is very difficult, and testing them is very difficult and

• 00:45:23 very difficult and complicated. Errors in self-renewal

• 00:45:26 complicated. Errors in self-renewal

• 00:45:31 can cause program overflows, but sometimes they make your job much easier. Does he know that

• 00:45:32 make your job much easier. Does he know that writing a self-referential program requires serious

• 00:45:35 writing a self-referential program requires serious experience.

• 00:45:39 experience. Interrupts can terminate a critical process

• 00:45:42 Interrupts can terminate a critical process and make a program

• 00:45:44 and make a program difficult to understand.

• 00:45:46 difficult to understand. A legacy of Hades Kot could not

• 00:45:51 A legacy of Hades Kot could not be placed. Why? Because

• 00:45:54 be placed. Why? Because you're making code in another class,

• 00:45:57 you're making code in another class, you're stuck. You say these

• 00:46:00 you're stuck. You say these changes were made, and unexpected

• 00:46:02 changes were made, and unexpected behavior can cause problems. I don't understand the code.

• 00:46:04 behavior can cause problems. I don't understand the code. So,

• 00:46:08 So, if a change is made to the pot you inherited, it can affect code in a very different class. You

• 00:46:10 if a change is made to the pot you inherited, it can affect code in a very different class. You can break

• 00:46:14 can break it. Love made by Ali Ay Sing.

• 00:46:18 it. Love made by Ali Ay Sing.

• 00:46:23 How can using multiple horses to refer to the same state variable in Turkish change this? This

• 00:46:37 means aliasing. Yes, no signal processing. Think about it. This conversation depends

• 00:46:39 Think about it. This conversation depends on the concept used. Unlimited arrays.

• 00:46:43 on the concept used. Unlimited arrays. If there is no bound control in arrays, buffer

• 00:46:46 If there is no bound control in arrays, buffer overflow errors can occur.

• 00:46:48 overflow errors can occur. Default input processing and input

• 00:46:51 Default input processing and input independence. Is an input action occurring in a

• 00:46:53 independence. Is an input action occurring in a default action degree transferring the program to

• 00:46:56 default action degree transferring the program to another location. This

• 00:47:00 another location. This can cause problems when a program restarts, which can trigger incorrect and

• 00:47:01 can cause problems when a program restarts, which can trigger incorrect and deliberately malicious

• 00:47:03 deliberately malicious program pain.

• 00:47:08 program pain.

• 00:47:17 For systems that provide capabilities and require long processes or user interaction, you should always start from

• 00:47:23 a starting point that allows users to restart everything they do without having to redo everything.

• 00:47:27 For example, if the power goes out and there is no power supply and the

• 00:47:31 goes out and there is no power supply and the process is very long,

• 00:47:34 process is very long, starting over from the beginning can be very

• 00:47:37 starting over from the beginning can be very time-consuming and costly.

• 00:47:41 time-consuming and costly. Of course, if an error

• 00:47:45 Of course, if an error occurs, a restart is necessary. This

• 00:47:48 occurs, a restart is necessary. This also has the advantage of

• 00:47:51 also has the advantage of being a restart system. This means that the system

• 00:47:53 being a restart system. This means that the system stores copies of the forms connected to the system.

• 00:47:56 stores copies of the forms connected to the system. This way,

• 00:47:58 This way, they don't have to refill the problem.

• 00:47:59 they don't have to refill the problem. For example, this is a type of periodic

• 00:48:03 For example, this is a type of periodic saving and restarting from the saved state.

• 00:48:05 saving and restarting from the saved state. This is a second situation.

• 00:48:10 This is a second situation. We check the boundaries of the

• 00:48:12 We check the boundaries of the array. Some programming languages, such as Erce,

• 00:48:15 array. Some programming languages, such as Erce, do not listen to the

• 00:48:17 do not listen to the array declaration. In some programming languages, an array

• 00:48:20 array declaration. In some programming languages, an array declaration is outside the allowed range. This

• 00:48:22 declaration is outside the allowed range. This address is outside the allowed range.

• 00:48:23 address is outside the allowed range. You are using the Republic today. This

• 00:48:29 You are using the Republic today. This attackers

• 00:48:32 attackers

• 00:48:37 can deliberately take the code beyond the best of us by writing it into memory. This

• 00:48:42 leads to a security vulnerability. If one of you doesn't include bounded checking, that is, your programming language.

• 00:48:46 doesn't include bounded checking, that is, your programming language. Therefore, always

• 00:48:48 Therefore, always checking whether an array is within the bounds of the array

• 00:48:49 checking whether an array is within the bounds of the array

• 00:48:55 can also lead to these vulnerabilities. When searching for external components, include the time portions of the external components.

• 00:48:58 include the time portions of the external components. In a distributed system,

• 00:49:01 In a distributed system, a remote computer failure can occur silently,

• 00:49:03 a remote computer failure can occur silently, without you ever knowing about it. Programs

• 00:49:05 without you ever knowing about it. Programs waiting for a service from that computer never receive

• 00:49:07 waiting for a service from that computer never receive that service or

• 00:49:09 that service or any indication of a failure.

• 00:49:11 any indication of a failure. So,

• 00:49:15 So, they can wait forever, not forever. To avoid this,

• 00:49:18 they can wait forever, not forever. To avoid this,

• 00:49:27 you can always open a time window for all calls made to external components. After a defined period of time has passed without a response, you can create

• 00:49:30 has passed without a response, you can create a web page that appears to be in front of

• 00:49:32 a web page that appears to be in front of you. If the result

• 00:49:34 you. If the result doesn't respond, you'll have to wait a few seconds.

• 00:49:36 doesn't respond, you'll have to wait a few seconds. If you're the one who assumes the failure is real, you'll need to consider the error.

• 00:49:39 If you're the one who assumes the failure is real, you'll need to consider the error.

• 00:49:43 Hello friends. So, what will the website say? If it

• 00:49:48 doesn't respond, I'll throw the correct Time at error. So, you open the error.

• 00:50:01 Name all constants that represent real-world values. Instead of using numerical values, always

• 00:50:04 Instead of using numerical values, always

• 00:50:09 give constants that reflect real-world values, such as tax rates. We always assign them by name and

• 00:50:11 always assign them by name and find them. When using at instead of value, you're less likely to

• 00:50:15 find them. When using at instead of value, you're less likely to make mistakes and avoid typing incorrect things.

• 00:50:16 make mistakes and avoid typing incorrect things. Especially

• 00:50:19 Especially if you have a compile-time check,

• 00:50:23 if you have a compile-time check, this greatly reduces

• 00:50:27 this greatly reduces the probability of error.

• 00:50:30 the probability of error. This is another advantage.

• 00:50:33 This is another advantage. Of course, these times are not truly constant. So, you

• 00:50:37 Of course, these times are not truly constant. So, you only

• 00:50:39 only need to make the change in one place in your program. Because

• 00:50:43 need to make the change in one place in your program. Because you always called that name, you changed it, your charge

• 00:50:45 you always called that name, you changed it, your charge named, instead of that variable containing ms-100,

• 00:50:47 named, instead of that variable containing ms-100, you didn't write a hundred. Look at it as a number. The

• 00:50:50 you didn't write a hundred. Look at it as a number. The key points for Hello are

• 00:50:53 key points for Hello are difficult to access software varieties. Because it's

• 00:50:55 difficult to access software varieties. Because it's

• 00:50:59 almost impossible to ensure that each version of the software is truly independent. For the reasons I've explained, it relies on the inclusion of

• 00:51:02 For the reasons I've explained, it relies on the inclusion of

• 00:51:09 too many got statements in a program to safely check the validity of programming inputs and the values ​​of program variables.

• 00:51:16 So, dive somewhere. Jump. Pointers, pointers, recursion,

• 00:51:19 Pointers, pointers, recursion, birkosit English, blind. You're left with

• 00:51:22 birkosit English, blind. You're left with

• 00:51:27 some programming structures and techniques like floating point, these simpoint numbers,

• 00:51:30 by nature, we made mistakes. In terms of the prayer back, whatever programming language

• 00:51:32 prayer back, whatever programming language you use,

• 00:51:34 you use,

• 00:51:40 you should try to avoid these structures when developing security systems. Yes, friends, let's create a fesi

• 00:51:43 create a fesi and hopefully you're right.

• 00:51:53 and hopefully you're right. We'll see you in the last lesson. You'll learn 14. We've

• 00:51:58 We'll see you in the last lesson. You'll learn 14. We've reached the end of our lesson.

• 00:52:02 reached the end of our lesson. Thank you for listening. Take care. Let's

• 00:52:45 Yes, friends. Hopefully, see you next week.