Hey All: I just started TDD TV. TDD Practitioners apply SOLID principles and all sorts of other practices like Clean Code, etc when we practice TDD. SUBSCRIBE & hit the BELL to Learn! Many videos to come! (Also check out WeDoTDD.com!) 👾📺👾📺👾📺👾📺👾📺👾📺👾📺👾📺👾📺👾📺👾📺👾📺👾📺👾📺👾📺
notes to self: 29:00 polymorphism is why OO is so successful 47:00 a DLL is not a logical grouping of classes but a grouping of classes that you want loaded at runtime 49:00 single responsibility means design the class so it has a single reason to change or, to put it in another way, gather the things that change for the same reasons and separate the things that change for different reasons 56:00 the database exists to make accessing data from a disk easier for us, disks are disappearing and in the future we will have virtually infinite amount of persistent ram 59:00 open-closed: being able to change what the class does without changing its source code (you write new code but don't touch old code)
So cool to hear him talking about the "old" days and how thing were actually implemented... I started with C and C++ and still remember lots of this stuff.
I'm starting to get why they call him "Uncle Bob". It's the crazy stories that make you think about the other more serious stuff he's saying. He oscillates wildly between foolishness and wisdom, which is the sign of a true teacher.
Wonderful insight into the history of OOP and what's it's really about. In fact, OO can be just an extension to some functional language or procedural one.. it has nothing to do with mutability or state at all.
James Gosling was at the language naming brainstorming session where people were drinking Peet's Java and someone suggested "Java", but no one remembers who it was. Some were dead set against Java as the name because they wanted descriptive names for it and a related product(s). Java and Silk may have been the only two names approved by legal at the time management insisted a name be chosen.
Uncle Bob says programmers choose the computers at their companies. I don't think that has ever been my experience. When IBM first started producing PCs, they gave them to the Mainframe Salespeople who gave them to CTOs as presents. The CTOs didn't know what to do with them and stuck them in a closet. My experience was that microcomputers were introduced into companies by CTOs to control peripherals that IBM and DEC would not allow them to connect to the mainframes and minicomputers, for branch networks in Banks, and Wall Street quants who didn't want their formulae on insecure mainframes. These were architectural decisions and had nothing to do with programming languages. I was at Sun during the dotcom boom. Unlike its competitors, Sun's computers were web-ready, configured with massive numbers of threads, etc. The Salespeople were selling hardware to CTOs and threw in Professional Services for free. Our ability to readily create scalable Client-Server systems using Java and JSPs landed a lot of sales, and further lucrative work.
At 20:05, Bob explains C had perfect encapsulation, and explains why C++ does not. He says, in C++, since the new operator needs to know the size of a class's object, it's definition therefore must be in a header file, hence losing encapsulation. My question is: Even in C, wouldn't you need to know the size of the structure you want to malloc?
That's a good question! I think he means the paradigm encouraged by C++, instead of what you can do with C++. In C, you'd often have a function that would do the malloc for you, and you'd be responsible for calling the corresponding clean up function once you were done. As long as you'd be operating with a pointer to the struct and pass that to the functions in the header, the compiler would allow you to do that. You can do this in C++ too, but I believe the point Bob is making is about going the C++ way, i.e. using classes and 'new'.
Ha. Yes. I came here specifically hoping for I. and D. since in the previous talked I just watched, he only got through S.O.L. Someone should try to book him for a D.I.L.O.S. talk.
You say that Sun was the first to realize that winning the hearts and minds of programmers. Way back in the 50s and 60s IBM figured it out too and gave universities huge discounts on systems. Unfortunately for IBM they forgot about it.
+Robert Leather I strongly disagree and I think you are missing the main point : AGILE. To design software systems in agile, Uncle Bob always suggests to think in the most loosely coupled ways of structuring the program blocks. Starting with the business logic with no gui and no db, which can be tested in each iteration, you will add your gui and db as dynamically "pluggable" modules. Most funniest-perfect way to tell the programmers that they should think beyond the injected and repeated keywords "inheritance, polymorphism etc and make things loosely coupled.
+Dave Schinkel Really appreciate your team effort to improve the coding and having such great sessions, which we could watch on-line. I started following Uncle Bob right when I found them from NDC at Vimeo. Since then, kept on watching such videos, then further debates and hangout sessions from David (DHH) , Beck and Martin. Started practicing it throughout my codings, refactoring (using Boy's scout rule) and in an year, the program I wrote became complex; yet well maintainable. I could change my UI from native to web without cascading it to my logic, or change the database without cascading again. First time I really felt; "yeah! I did something good". What I missed is adding unit tests from the beginning; which I am trying to incorporate now. :-) . Problem again is the mindset from the management on spending that 30 seconds extra for unit tests and further the TDD :( Not yet there when marketing guys can just take the latest build and sell it when they want it, but I am striving to.
+Dave Schinkel Absolutely Dave! I had a lovely day programming thanks to using TDD. All is covered, everything in the green, great tangible process. Now I have to stop working, but I don't want to.. so what better alternative than watching Uncle Bob :) I'll eat bread & water before dealing with fast, messy and untrustworthy code again!
57:18 I don't think tables (arrays) are going away any time soon. It's true we don't have spinning disks anymore, but it still is much more efficient to read data that is contiguously laid out. Also, even though we'll have more memory in the future we'll still need to serialize our data. And lastly, computers aren't getting that much faster anymore. All capacities are still finite, and the idea that we don't need to care will (and has) lead to poor performance, and as such, poor software quality. In the end, programming is engineering, and engineering is about the physical reality. I don't think trying to 'escape the hardware' as an underlying principle and goal is a wise way forward. Abstractions aren't like math, because all abstractions are imperfect and each one comes with a complicated price tag. I absolutely agree with Bob on the point he makes about not "trusting" frameworks. The technical relationship is the wrong way around. Most websites have somehow missed this point, but then again web development is very broken.
The moon is made of the same material as the earth, except for the iron core of the earth, which comes from another body (the body of the size of Mars), but it couldn't belong to the solar system, because another sun must have produced it, the solar system either captured or it simply collided with earth.
A correction about similarities between Java and C#. While they are similar, some of the similarities are only on the surface. Yes, the syntax is the same but the implementation is different.
This whole talk is about how to communicate with your code, to OTHER PROGRAMMERS. Your point is literally, that machines read 100101010101010 NOT code. So the point of this talk is -> How to write code a human can understand. And your correction is, to say, what was already said... You weren't paying attention my friend.
The comment made by the audience suggests that uninstalling Resharper(Jetbrains) from your Windows(Microsoft) computer is not easy, and that it's a headache(it's difficult and unpleasant) to uninstall.
I think it implies that if user uninstall re-sharper Microsoft will be in trouble because v. studio won't be that effective without it. So Jetbrains do have a way to damage microsoft
I watched almost all of them and I highly recommend them. Although the Design Patterns section of it wasn't very helpful but the rest specially SOLID section was fantastic.
He never really covers them because they arent as interesting and as insidious and subtle as the first 3 I feel. You can just read them from his book. For example DIP is simply to use an interface because high level (ie policy) modules shouldn't contain direct references (be dependent) on particular implementation. Its not much more nuanced than that. But the points he covers are way more nuanced.
Dave Schinkel thanks! I got the second part of the question which is about dependency injection or inversion. Couldn't catch the first part that was answered with the rect/square example.
6 ปีที่แล้ว +1
Oh, FYI, just got what they are talking about by paying attention to Uncle Bob repeating the question. He says: "Should i touch briefly on LISKOV----- and Dependency Inversion?". Those are the L and the D from SOLID. That's why he talks about the rectangle/square example, and why square cannot inherit from rectangle. Basically the answer is: "both Rectangle and Square should inherit from SHAPE, to comply with the Liskov's substitution principle". A Square IS a Rectangle, and that's Uncle Bob's bait because Square doesn't BEHAVE as a rectangle. en.wikipedia.org/wiki/Liskov_substitution_principle
@@ninjedi6710I’m enjoying miltiparadigmal languages. Even Java now have functional features so I’m not the only one. Some tasks are cleaner in fp, some in oop. Very nice when language allows combining
Hey All: I just started TDD TV. TDD Practitioners apply SOLID principles and all sorts of other practices like Clean Code, etc when we practice TDD.
SUBSCRIBE & hit the BELL to Learn! Many videos to come! (Also check out WeDoTDD.com!)
👾📺👾📺👾📺👾📺👾📺👾📺👾📺👾📺👾📺👾📺👾📺👾📺👾📺👾📺👾📺
notes to self:
29:00 polymorphism is why OO is so successful
47:00 a DLL is not a logical grouping of classes but a grouping of classes that you want loaded at runtime
49:00 single responsibility means design the class so it has a single reason to change or, to put it in another way, gather the things that change for the same reasons and separate the things that change for different reasons
56:00 the database exists to make accessing data from a disk easier for us, disks are disappearing and in the future we will have virtually infinite amount of persistent ram
59:00 open-closed: being able to change what the class does without changing its source code (you write new code but don't touch old code)
Before the talk starts: "Ok what crazy tangent is he going to start with this time?"
Talk starts: "WHY DO WE HAVE A MOON?"
Not disappointed lmao
I could listen to him for hours.
I have been, and I can tell you, it is worth it
All talks of him I find on youtube are roughly the same though :(
I do same
Most I’ve learned from a single TH-cam video in a very long time! Thank you for posting this lecture!
Gotta love Bob's little tangents he has before he starts his talks. It is really effective at getting the listener's engaged in the talk.
I wish I could upvote this more than once. Take a bow!!!
So cool to hear him talking about the "old" days and how thing were actually implemented... I started with C and C++ and still remember lots of this stuff.
I'm starting to get why they call him "Uncle Bob". It's the crazy stories that make you think about the other more serious stuff he's saying. He oscillates wildly between foolishness and wisdom, which is the sign of a true teacher.
I thought the name came from a co-worker who liked giving odd names to their colleagues.
To me he is the Stephen King of code. Always on the edge of your seat.
And the sign of a true uncle.
A fascinating overview of the history of programming languages. After 50 mins two SOLID principals are introduced SRP and OCP but only surface deep.
I haven't seen 1 lecture of him explaining the whole 5 principles
I did clean code course with uncle bob and the tangent talks at start of each topic kept you focused also they were really interesting stories
Wonderful insight into the history of OOP and what's it's really about. In fact, OO can be just an extension to some functional language or procedural one.. it has nothing to do with mutability or state at all.
9:30 "We used to call it C++ for dummies."
James Gosling was at the language naming brainstorming session where people were drinking Peet's Java and someone suggested "Java", but no one remembers who it was.
Some were dead set against Java as the name because they wanted descriptive names for it and a related product(s).
Java and Silk may have been the only two names approved by legal at the time management insisted a name be chosen.
Uncle Bob is a national treasure.
starts at 3:23
Anecdote ends at 8:14
doing god's work
What a fun and engaging speaker! Thanks.
Uncle Bob says programmers choose the computers at their companies. I don't think that has ever been my experience.
When IBM first started producing PCs, they gave them to the Mainframe Salespeople who gave them to CTOs as presents. The CTOs didn't know what to do with them and stuck them in a closet.
My experience was that microcomputers were introduced into companies by CTOs to control peripherals that IBM and DEC would not allow them to connect to the mainframes and minicomputers, for branch networks in Banks, and Wall Street quants who didn't want their formulae on insecure mainframes. These were architectural decisions and had nothing to do with programming languages.
I was at Sun during the dotcom boom. Unlike its competitors, Sun's computers were web-ready, configured with massive numbers of threads, etc. The Salespeople were selling hardware to CTOs and threw in Professional Services for free. Our ability to readily create scalable Client-Server systems using Java and JSPs landed a lot of sales, and further lucrative work.
At 20:05, Bob explains C had perfect encapsulation, and explains why C++ does not. He says, in C++, since the new operator needs to know the size of a class's object, it's definition therefore must be in a header file, hence losing encapsulation. My question is: Even in C, wouldn't you need to know the size of the structure you want to malloc?
That's a good question! I think he means the paradigm encouraged by C++, instead of what you can do with C++. In C, you'd often have a function that would do the malloc for you, and you'd be responsible for calling the corresponding clean up function once you were done. As long as you'd be operating with a pointer to the struct and pass that to the functions in the header, the compiler would allow you to do that. You can do this in C++ too, but I believe the point Bob is making is about going the C++ way, i.e. using classes and 'new'.
Great talk, as usual with Uncle Bob. Too bad he only made it to the "S.O." out of "S.O.L.I.D.".
Ha. Yes. I came here specifically hoping for I. and D. since in the previous talked I just watched, he only got through S.O.L. Someone should try to book him for a D.I.L.O.S. talk.
This guy is crazy, but I really like his videos
You say that Sun was the first to realize that winning the hearts and minds of programmers. Way back in the 50s and 60s IBM figured it out too and gave universities huge discounts on systems. Unfortunately for IBM they forgot about it.
This guy is awesome.
14:26 Algol: "Algorithmic Language" also a star and Arabic for Daemon (al ghoul)
this guy is golden
So wait how do we fix the class he described in the Single responsibility principle example?
Edit : Oh ok he talks about it at 1:05:00
I loved this video, but lets be honest... all those wonderful recollections meant we didn't actually cover the subject of the video.
+Robert Leather
I strongly disagree and I think you are missing the main point : AGILE. To design software systems in agile, Uncle Bob always suggests to think in the most loosely coupled ways of structuring the program blocks. Starting with the business logic with no gui and no db, which can be tested in each iteration, you will add your gui and db as dynamically "pluggable" modules.
Most funniest-perfect way to tell the programmers that they should think beyond the injected and repeated keywords "inheritance, polymorphism etc and make things loosely coupled.
+Dave Schinkel
Really appreciate your team effort to improve the coding and having such great sessions, which we could watch on-line.
I started following Uncle Bob right when I found them from NDC at Vimeo. Since then, kept on watching such videos, then further debates and hangout sessions from David (DHH) , Beck and Martin. Started practicing it throughout my codings, refactoring (using Boy's scout rule) and in an year, the program I wrote became complex; yet well maintainable. I could change my UI from native to web without cascading it to my logic, or change the database without cascading again. First time I really felt; "yeah! I did something good".
What I missed is adding unit tests from the beginning; which I am trying to incorporate now. :-) . Problem again is the mindset from the management on spending that 30 seconds extra for unit tests and further the TDD :(
Not yet there when marketing guys can just take the latest build and sell it when they want it, but I am striving to.
+Dave Schinkel Absolutely Dave! I had a lovely day programming thanks to using TDD. All is covered, everything in the green, great tangible process. Now I have to stop working, but I don't want to.. so what better alternative than watching Uncle Bob :)
I'll eat bread & water before dealing with fast, messy and untrustworthy code again!
57:18 I don't think tables (arrays) are going away any time soon. It's true we don't have spinning disks anymore, but it still is much more efficient to read data that is contiguously laid out. Also, even though we'll have more memory in the future we'll still need to serialize our data. And lastly, computers aren't getting that much faster anymore. All capacities are still finite, and the idea that we don't need to care will (and has) lead to poor performance, and as such, poor software quality. In the end, programming is engineering, and engineering is about the physical reality. I don't think trying to 'escape the hardware' as an underlying principle and goal is a wise way forward. Abstractions aren't like math, because all abstractions are imperfect and each one comes with a complicated price tag.
I absolutely agree with Bob on the point he makes about not "trusting" frameworks. The technical relationship is the wrong way around. Most websites have somehow missed this point, but then again web development is very broken.
Thanks for the upload, a really fascinating talk.
I love this man
Does he have a problem with C# ?
Is there some passive-aggression in his tone? Not sure!
What IDE does he name at ~ 35:18?
I can't make it out & I don't work with C#, NET, or Java.
Many Thanks. 8)
Uncle Bob
8:15 where the actual talk starts
The moon is made of the same material as the earth, except for the iron core of the earth, which comes from another body (the body of the size of Mars), but it couldn't belong to the solar system, because another sun must have produced it, the solar system either captured or it simply collided with earth.
Sorry it's my audio device.
Doesn't PL/SQL have functions local to other functions?
It is the fifth not the 2nd or 3rd.
18:36 bob gettin angery
at 36:50 what did the guy say?
A correction about similarities between Java and C#.
While they are similar, some of the similarities are only on the surface.
Yes, the syntax is the same but the implementation is different.
This whole talk is about how to communicate with your code, to OTHER PROGRAMMERS.
Your point is literally, that machines read 100101010101010 NOT code.
So the point of this talk is -> How to write code a human can understand.
And your correction is, to say, what was already said...
You weren't paying attention my friend.
SOLID at th-cam.com/video/t86v3N4OshQ/w-d-xo.htmlm16s
Can someone explain joke from 36:50 to not native English speaker please?
The comment made by the audience suggests that uninstalling Resharper(Jetbrains) from your Windows(Microsoft) computer is not easy, and that it's a headache(it's difficult and unpleasant) to uninstall.
I think it implies that if user uninstall re-sharper Microsoft will be in trouble because v. studio won't be that effective without it.
So Jetbrains do have a way to damage microsoft
I think he says "have you ever tried to uninstall it?"
Because it’s sometimes even when you uninstall it, it won’t get completely uninstalled😂
Three principles are covered nicely, but last two are not.
naresh dhami its always the same, i cannot find a talk about everything
+naresh dhami You can go to his own website and watch more videos like this. www.cleancoders.com
I watched almost all of them and I highly recommend them. Although the Design Patterns section of it wasn't very helpful but the rest specially SOLID section was fantastic.
He never really covers them because they arent as interesting and as insidious and subtle as the first 3 I feel. You can just read them from his book. For example DIP is simply to use an interface because high level (ie policy) modules shouldn't contain direct references (be dependent) on particular implementation. Its not much more nuanced than that.
But the points he covers are way more nuanced.
does anyone have a link where he gets past o?
CleanCoders.com
C# is Java done right;-)
Microsoft products are not for programmers
Java is also good. Java has much better backwards compatibility, in c# we had great framework/core split as well as many other breaking changes.
did he mention objective c and jobs?
Not in this talk, he mentioned Java and C# instead. In one of his other SOLID talks, he talked about Objective C and Steve Jobs.
Don't get screwed by your GUI
Who else is doing ICT283?
What's the question in 1:08:35 ?
Dave Schinkel thanks! I got the second part of the question which is about dependency injection or inversion. Couldn't catch the first part that was answered with the rect/square example.
Oh, FYI, just got what they are talking about by paying attention to Uncle Bob repeating the question. He says: "Should i touch briefly on LISKOV----- and Dependency Inversion?". Those are the L and the D from SOLID.
That's why he talks about the rectangle/square example, and why square cannot inherit from rectangle. Basically the answer is: "both Rectangle and Square should inherit from SHAPE, to comply with the Liskov's substitution principle".
A Square IS a Rectangle, and that's Uncle Bob's bait because Square doesn't BEHAVE as a rectangle.
en.wikipedia.org/wiki/Liskov_substitution_principle
Uncle Bob entere 3:20
S.O.L.I.D. principles start at 49:18
OOP > FP
You haven’t done FP have you. And if you did it was a JVM based language wasn’t it.
@@HikarusVibrator I work with react. FP is my daily life.
@@HikarusVibrator and python for backend
@@ninjedi6710I’m enjoying miltiparadigmal languages. Even Java now have functional features so I’m not the only one. Some tasks are cleaner in fp, some in oop. Very nice when language allows combining
😮