Erik Meijer, Gilad Bracha, Mads Torgersen: Perspectives on Programming Language Design and Evolution

cool too bad gilad has to be so gosh darned arrogant.. everything he didnt make seems to suck/be evil/sick and every one who doesnt agree with him he seems to think is an idiot:P he has alot of valid points but i just dont think the world is that black and white..

but hey.. thats just my impression awsome interview

Great video!  I have been debating the pros and cons of reflective programming with somebody in the Techoff threads who is adamantly against it, and many of the points I would have liked to have made in favor of dynamic languages and better type systems were made here in this video and in a more eloquent manner.  Keep up the good talks, Charles!

Charles wrote:


Gilad has a complex sense of humor.

indeed

are there any more lang.net vids coming up btw?

Thx Charles for awesome interviews, its so Nice to hear such kind of talks on programming languages...and the future...

I hope we have some more from Lang.NET 2008..

Big Cheers...

I have to agree... I found this talk interesting but Gilad comes across as very arrogant. He seems to think that most developers don't want to learn anything. If that were the case no one would be watching this video.

I think perhaps working in an academic scenario for a long time causes people to lose appreciation for people working against commercial pressures. Most developers I have worked with are very keen to learn new languages/technologies but in reality they have deadlines to meet. In addition it takes time for these things to gain acceptance at the corporate level.

His comments about older people's inability to learn are also out of line.

It's a shame really; I feel these comments tarnished an otherwise very interesting discussion.

If the session Gilad held on newspeak is the same as he did on Jaoo, thats some extremely crazy stuff.

Good interview. I disagree with the people who say that Gilad comes out arrogant.

Charles wrote:

I have more from Lang.NET in the pipeline. Stay tuned.
C

Is there gonna be more with Mads? He seems extremely cool, but he didnt get to say all that much compared to Erik and Gilad in this interview, which is a shame.

Great...Charles..
I will enjoy all the stuff from Lang.NET 2008

Also, Gilad had his viewpoints and is nothing wrong or arrogant..
Awesome video..

Just great video,

So much thanks go to all of you out there at Channel 9 for getting so much great material on my screen’s pixel.

I just love your coverage about functional programming and the talks you can have with all those gurus.

Thanks again for giving us the chance to feel ourselves on the campus or on the conferences with you.  You fell less alone isolated in your “little geekless towns”.

Keep up the great work !  I know I will be watching

One of the best videos in a long time and a bit of an eye-opener. The art of molding software and logic. Also - this looks highly interesting (bookmarked!) http://gbracha.blogspot.com/

That's exactly it. Cheers, Gilad

This is great stuff. I think the idea of a pluggable type system is interesting(seems cleaner to me), but would it really add to the power of a type system say of the CLR.

My meaning here is, at some level all types are comprised of some other basic types. For example, you could say have a variable of type bmp... but at a basic level it can be expressed as a byte array.

Unless we start doing things like var x  = new SQLserver(), wouldn't a pluggable type system just allow us to more easily work with this super set of types? But then again, couldn't the same be accomplished by creating the class and doing a lot of operator overloading? Or, would this be a way to dynamically type these complex object types?

Or, and this very well could be the case, am I just missing the point?


edit: just an after thought. its mentioned that a built in type system can limit the ability of a language to adapt over time... but the first thing that pops to mind is... Fortran had ints an strings and arrays... and so does the CLR(albeit a little smarter). Theres nearly 50 years between now and when fortran came out.. Just for a chuckle.

Dark_Halmut wrote:

alot of stuff.

I didnt really read your post. But your avatar is hysterical. I have thought about using that picture myself..

But it isnt exactly.. Politically correct:(

gaurav.net wrote:


Also, Gilad had his viewpoints and is nothing wrong or arrogant..

well ofcourse he is entitled to his own oppinion
but i think he is definitly wrong about people not wanting to learn.. its very hard not to learn as a programmer.. its a creative process. i think that if you dont like learning you will fail as a programmer and end up doing other work. also, as previously stated, real world programmers does have a job to do. they dont always have the time to stroke their beards and bury themselvs in theory, but that doesnt mean they dont want to, i whould certinaly like to have time to do that (as long as its useful atleast).

what i find arrogant is his blatant disses of far more successful technoligis like c#, java, the clr and c++ as crap (compared to newspeak and smalltalk), and that he attribute their success to programmer ignorance rather than quality of the respective systems..

its like its never the fault of smalltalk or erlang or haskell or whatever that they are not widly accepted, because they have no flaws (according to their inventors/proponents). its just that everone else is to stupid to understand their greatness. i know im exaggerating but you get my point

to me, thats the same kind of rethoric that mac and linux fanbois (that do not make up the entirety of those communities) use against windows. "its crap and people only use it because they're stupid, if they where smart (like me) the whould use the system im using"

i know he doesnt say that but to me, he(and many others) comes across like meaning that..  it doesnt really sound like hes making a joke, but charles tells me his sense of humor is complex..  i think "jokes" like are very ofputting to the programmer comunities he whould want to use stuff like newspeak. if youre reading this gilad, please consider this constructive feedback.

a great example of someone who does not come across as arrogant and make "his" technique sound very appealing and approachable is erik meijer. he does a great job explaining the concepts of fp without make it uncessecary complicated. so major bigup on erik. i think Gilad whould have a far easier time getting smalltalk and newspeak widely accepted if he took a page from erik

and another thing, saying that you can make dynamic systems fast and efficient is not good enough. you have to actually do it. like what microsoft has done with the dlr.. twice as fast as cpython last year, and its built on the clr with its static type system

but again, a very interesting video, just because i disagree with some stuff doesnt make it bad

I think Gilad was *absolutely spot on* about moving the industry forward and somebody is bound to feel "hit" by what he's saying, but it's not arrogance, it's just stating how things are (or at the very least an oppinion of how things are, agree with it or not)... I really enjoy his sense of humor and point of view.

aL_ wrote:



gaurav.net wrote:  Also, Gilad had his viewpoints and is nothing wrong or arrogant..

well ofcourse he is entitled to his own oppinion
but i think he is definitly wrong about people not wanting to learn.. its very hard not to learn as a programmer.. its a creative process. i think that if you dont like learning you will fail as a programmer and end up doing other work. also, as previously stated, real world programmers does have a job to do. they dont always have the time to stroke their beards and bury themselvs in theory, but that doesnt mean they dont want to, i whould certinaly like to have time to do that (as long as its useful atleast).

what i find arrogant is his blatant disses of far more successful technoligis like c#, java, the clr and c++ as crap (compared to newspeak and smalltalk), and that he attribute their success to programmer ignorance rather than quality of the respective systems..

its like its never the fault of smalltalk or erlang or haskell or whatever that they are not widly accepted, because they have no flaws (according to their inventors/proponents). its just that everone else is to stupid to understand their greatness. i know im exaggerating but you get my point

to me, thats the same kind of rethoric that mac and linux fanbois (that do not make up the entirety of those communities) use against windows. "its crap and people only use it because they're stupid, if they where smart (like me) the whould use the system im using"

i know he doesnt say that but to me, he(and many others) comes across like meaning that..  it doesnt really sound like hes making a joke, but charles tells me his sense of humor is complex..  i think "jokes" like are very ofputting to the programmer comunities he whould want to use stuff like newspeak. if youre reading this gilad, please consider this constructive feedback.

a great example of someone who does not come across as arrogant and make "his" technique sound very appealing and approachable is erik meijer. he does a great job explaining the concepts of fp without make it uncessecary complicated. so major bigup on erik. i think Gilad whould have a far easier time getting smalltalk and newspeak widely accepted if he took a page from erik

and another thing, saying that you can make dynamic systems fast and efficient is not good enough. you have to actually do it. like what microsoft has done with the dlr.. twice as fast as cpython last year, and its built on the clr with its static type system

but again, a very interesting video, just because i disagree with some stuff doesnt make it bad

I Completely agree with you .., with every single line indeed, specially that about the windows-linux thing...
Erik Meijer is great, & he makes us feel even more great that is what the most important part is...
DLR seems quite interesting to me now..., it will play a major role in future versions of .NET...
Once again thx to Charles for such an awesome video...

Charles does he have any with Anders Hejlsberg...?
Cheers for this one, it was awesome and great..

esoteric wrote:

I think Gilad was *absolutely spot on* about moving the industry forward and somebody is bound to feel "hit" by what he's saying, but it's not arrogance, it's just stating how things are (or at the very least an oppinion of how things are, agree with it or not)... I really enjoy his sense of humor and point of view.

i agree that the industry should move forward but im not sure smalltalk/newspeak is that way forward.. i guess i require more concrete stuff, not comments about how bad everything else is
 
i deffinietly do not agree that things are they way he says they are (that c/c++ is evil and that the clr is funamentally flawed among other things). i dont think the clr is perfect but its not worthless either.  making highly categorical claims like that does not make one more convincing in my book (rather the opposite).

but again, he is ofcourse entiteld to his oppinion, but  in order for me to acept his view and correct i need more proof

show me a completly virtual/dynamic vm that is  "better" (more usefull and atleast as fast) than the clr and i'll gladly use it, but until then i'll stick with the clr   again, just saying "its possible" is not good enough for the real world programmer, you gotta put your money where your mouth is, so to speak

Thank you to all in the interview, and the interviewer. Gilad is not arrogant. His comments are right on target. He does not avoid strong opinions.

These videos reach a much larger audience than the people at a programming language conference. While some may react badly to criticisim of their favorite technology (i.e. severe limitations in the CLR), the videos are a wonderful way to spread ideas to the larger audience. It is a great way to do "marketing" for these ideas.

"Drivers are easy"

Also, just a thought... In the boring world of business programming we like rigidity and rules.

The Singualrity OS proved that inter-process communication, at least, can be faster when using a higher level language, as long as strict contracts are observed. This means drivers are best not written in a language such as c++, since they are a critical part of the OS stack.

Charles wrote:

In Gilad's defense, I should stress that this was a conversation on perspectives (and therefore opinions) on the history, present and future of general purpose programing languages and associated runtime environments. Like all of us, Gilad is entitled to his opinion and if you disagree with his viewpoints, well, that's fine. Erik defended C++ (as I did), but Gilad has a point here. The fact that a language is successful does not mean that the language, from purely a language perspective, is successful.

I think it's awesome to showcase multiple viewpoints on a big topic like this. Gilad is not arrogant. He's strongly opinionated.  This is a very good thing for the field.

you know gilad far better than me so i'll take your word for that
im just saying, thats how he comes across to me in this interview(and others aswell)
other niners disagree and thats fine

it's true that there are diffrent views of what makes a language successful, me beeing a industrial kind of programmer i guess i like pragmatism if i was a language designer, i can see how my views could be diffrent, like linguistic "beauty"

Frank Hileman wrote:

Gilad is not arrogant. His comments are right on target. He does not avoid strong opinions.

These videos reach a much larger audience than the people at a programming language conference. While some may react badly to criticisim of their favorite technology (i.e. severe limitations in the CLR), the videos are a wonderful way to spread ideas to the larger audience. It is a great way to do "marketing" for these ideas.

one can have strong oppinions without coming across as arrogant
whether gilad "is" arrogant or not, in truth, i cannot say, because i dont know him. but to me some of his comments on other technologies are unfair and exaggerated

i consider unfairness and exaggerated critisism against "other" techs  arrogant, but that seems so stir up alot of feelings, so perhaps i should rephrease and say that he was unfair to those technologies instead.

FYI neither java or c++ are my "favorite technology" but i still think gilad was unfair to them. (i also think that while the clr isnt perfect, its unfair to say that the its "severly limited")

but again, as has been stated many times in this thread, all are entitled to their own opinion

Charles wrote:


My point was also that your view is yours and Gilad's is Gilad's. Neither is the "right" one. They just "are".

indeed

Charles wrote:


Keep on thinking!

will do

where there any more vids from lang.net coming btw? i know i asked before but im not sure i got an awnser.. but i might not have looked properly

-edit-
no i didnt look properly, i see you replied on another post and said there is more lang.net stuff coming up.
good times

-edit2-
hey anotherone is already up :O

Frank Hileman wrote:

The Singualrity OS proved that inter-process communication, at least, can be faster when using a higher level language, as long as strict contracts are observed. This means drivers are best not written in a language such as c++, since they are a critical part of the OS stack.

Depends. If your operating system trusts the people who are writing the drivers (as has tended to be the case in the past) then languages such as C++ and C will almost invariably be at least as fast as anything written in C# or other high-level pointerless languages.

I didn't find Gilad arrogant at all. I did find him to be Blunt and to the point.

Most in the language design world have pointed out the weaknesses in the CLR design. Even MS has acknowledged some of them. And it has been pointed out in earlier Video's.

Even C# was created as a better Java than Java.

The only way a Language improves is when the weakness is pointed out. sometimes the only way that is found is from the outside looking in.

On that point Newspeak is interesting with the "lack of a type system". and here is why.

A formalized type system can never module all the domains all the time.

To further that statement. If the type system in the framework was more flexable we wouldn't have to worry about the different domains that Linq is trying to solve. 

On the other matter it does solve most of the common problems with pain points when interfacing into other domains.

douglas

aL_ wrote:

(i also think that while the clr isnt perfect, its unfair to say that the its "severly limited")

From the perspective of someone implementing a language on a virtual machine, the CLR, with its baked in notions of types and polymorphism, is more limited than a simpler VM adding only things like garbage collection and object encapsulation, leaving dispatch and other details to the language implementer. From that perspective the CLR is limited. MSIL is essentially C# in a lower level form, not a general purpose VM.

Severely limited from the language implementer point of view, hence the DLR and other VMs since the CLR and JVM.

Charles,

It is really nice to see you guys start paying attention to environment, lighthing and sound is definitely is a step up.

It doesn't apply to this interview but a suggestion:
 
When guests are doing a screen demo, it would be much better if you can just capture the screen into a video rather than pointing the camera to it...

But so far so good. Thanks..

Emre

I didn't think Gilad was arrogant - he just had a very dry sense of humour. And like many that are (to use the American term) "super-intelligent" his "social" skills are perhaps not as developed as they are for most of us who don't have a fraction of his ability or talent.

Good video - beautifully controlled by Charles. A weaker interviewer would have struggled to get as much stuff out of his interviewees or would have taken offence at the perceived "arrogance" on display.

Thanks.

evildictaitor wrote:



Frank Hileman wrote: The Singualrity OS proved that inter-process communication, at least, can be faster when using a higher level language, as long as strict contracts are observed. This means drivers are best not written in a language such as c++, since they are a critical part of the OS stack.

Depends. If your operating system trusts the people who are writing the drivers (as has tended to be the case in the past) then languages such as C++ and C will almost invariably be at least as fast as anything written in C# or other high-level pointerless languages.

Correct. But OS hangups and crashes do not promote trust among end users. The interesting thing about higher level languages and formal methods is the promise of static analysis to both improve reliablity and increase performance. Right now the emphasis is on safety, not performance, but that could change.

I have finally had a chance to see this video. Given some comments, I expected that it would not be good at all but to me, Gilad is damn funny. He is sarcastic in a subtle way - my kind of guy and then throw erik in there, and well its a comedy club for nerds by nerds. Respect all around, I wish I could work with any of those guys. They are some of the best out there and I enjoyed this very much. I hope to learn whatever I can about Newspeak and as soon as I am done F#, I am going to hit Haskell and Erlang. Also thanks to Erik and Beckman, it seems I need to learn VB as well. But I have at least one trait as those guys - I love to learn. Time for some more coffee...

it's important to realize that just because we enjoy learning doesn't mean everyone else does.  it might even sound preposterous that there's people out there who dislike learning... in which case you have a lot to learn