Georg Wrede wrote:
> 
> Now, D is IMHO an excellent First Language. And some people really seem to find debuggers useful. Personally I never use them, because I find myself distracted and concentrating on the debugger instead of thinking of the problem.
> 
> With D, let folks have debuggers if they want. But to become a real programmer, one should learn to bike without the baby wheels.

Debuggers become quite useful in team environments where you didn't necessarily write a lot of the code in the program.  Particularly if it's an old program where DBC and similar approaches simply aren't feasible because the app was designed with the expectation that invalid parameters should be swallowed.  Sure I could still work out in my head how the thing is working, but if the system is alreadying held together by spit and bubblegum it may take quite a while to accomplish.  I think Linus' comments are valid for some portion of the software community, but not necessarily all of it.


Sean

In article <dtp0m9$5gl$1@digitaldaemon.com>, Sean Kelly says...
>
>Georg Wrede wrote:
>> 
>> Now, D is IMHO an excellent First Language. And some people really seem to find debuggers useful. Personally I never use them, because I find myself distracted and concentrating on the debugger instead of thinking of the problem.
>> 
>> With D, let folks have debuggers if they want. But to become a real programmer, one should learn to bike without the baby wheels.
>
>Debuggers become quite useful in team environments where you didn't necessarily write a lot of the code in the program.  Particularly if it's an old program where DBC and similar approaches simply aren't feasible because the app was designed with the expectation that invalid parameters should be swallowed.  Sure I could still work out in my head how the thing is working, but if the system is alreadying held together by spit and bubblegum it may take quite a while to accomplish.  I think Linus' comments are valid for some portion of the software community, but not necessarily all of it.
>
>
>Sean


I agree, Sean.  But looking over the article again, I have to observe that Linus was quite specific about his opinion as it applied to /kernel/ debuggers.  He doesn't seem to have given an opinion on the use of debuggers in general.  Thus, I may have overeacted about his opinion, though I still disliked his manner (I haven't the experience to counter his opinion about kernel debuggers).

The generalization occured as a result of Georg Wrede's post in which he took Linus opinion and expanded it into a distaste for debuggers in general.

-JJR

John Reimer wrote:
> In article <dtp0m9$5gl$1@digitaldaemon.com>, Sean Kelly says...
>> Georg Wrede wrote:
>>> Now, D is IMHO an excellent First Language. And some people really seem to find debuggers useful. Personally I never use them, because I find myself distracted and concentrating on the debugger instead of thinking of the problem.
>>>
>>> With D, let folks have debuggers if they want. But to become a real programmer, one should learn to bike without the baby wheels.
>> Debuggers become quite useful in team environments where you didn't necessarily write a lot of the code in the program.  Particularly if it's an old program where DBC and similar approaches simply aren't feasible because the app was designed with the expectation that invalid parameters should be swallowed.  Sure I could still work out in my head how the thing is working, but if the system is alreadying held together by spit and bubblegum it may take quite a while to accomplish.  I think Linus' comments are valid for some portion of the software community, but not necessarily all of it.
> 
> 
> I agree, Sean.  But looking over the article again, I have to observe that Linus
> was quite specific about his opinion as it applied to /kernel/ debuggers.  He
> doesn't seem to have given an opinion on the use of debuggers in general.  Thus,
> I may have overeacted about his opinion, though I still disliked his manner (I
> haven't the experience to counter his opinion about kernel debuggers).

I realized that later as well.  But I still disagree with Linus.  It seems like he's creating an artificially high barrier for entry in hopes that it will filter out some of the people he'd rather not work with, and to allow him to blame the inherent difficulty of kernel programming when someone requests a feature he doesn't want to implement.  But Linus has both the right to choose or reject whomever he wants as a co-developer, and the right to say "no" to anyone he chooses.  Upon reflection, it seems like Linus has simply constructed a situation that allows him to avoid confrontation and to make him feel like he's doing something really special.  He's welcome to do that, but I hope people realize what's actually going on when his bug fixes take five times longer than they need to.

> The generalization occured as a result of Georg Wrede's post in which he took
> Linus opinion and expanded it into a distaste for debuggers in general.

Aye.  "Real men don't use debuggers."  Personally, I can't relate to Georg's comment about how not using a debugger forces you to think about the problem in a different way.  But then I don't use the bulk of the fancier features provided by debuggers these days either, so perhaps that's part of the reason.  And with the state of things in D, I haven't used a debugger here at all :-p


Sean

Sean Kelly wrote:
> John Reimer wrote:
>> Sean Kelly says...
>>> Georg Wrede wrote:
>>> 
>>>> Now, D is IMHO an excellent First Language. And some people
>>>> really seem to find debuggers useful. Personally I never use
>>>> them, because I find myself distracted and concentrating on the
>>>> debugger instead of thinking of the problem.
>>>> 
>>>> With D, let folks have debuggers if they want. But to become a
>>>> real programmer, one should learn to bike without the baby
>>>> wheels.
>>> 
>>> Debuggers become quite useful in team environments where you
>>> didn't necessarily write a lot of the code in the program.
>>> Particularly if it's an old program where DBC and similar
>>> approaches simply aren't feasible because the app was designed
>>> with the expectation that invalid parameters should be swallowed.
>>> Sure I could still work out in my head how the thing is working,
>>> but if the system is alreadying held together by spit and
>>> bubblegum it may take quite a while to accomplish.  I think
>>> Linus' comments are valid for some portion of the software
>>> community, but not necessarily all of it.
>> 
>> I agree, Sean.  But looking over the article again, I have to
>> observe that Linus was quite specific about his opinion as it
>> applied to /kernel/ debuggers.  He doesn't seem to have given an
>> opinion on the use of debuggers in general.  Thus, I may have
>> overeacted about his opinion, though I still disliked his manner (I
>> haven't the experience to counter his opinion about kernel
>> debuggers).
> 
> I realized that later as well.  But I still disagree with Linus.  It
> seems like he's creating an artificially high barrier for entry in
> hopes that it will filter out some of the people he'd rather not work
> with, and to allow him to blame the inherent difficulty of kernel
> programming when someone requests a feature he doesn't want to
> implement.

Ever since before he went to the net about writing the OS, he was surrounded by a lot of friendly guys, not all of whom [how should I put this] were too comfortable with deep abstractions or hard-core concurrency issues.

Problem was, from day one, that these guys wanted so bad to help, but the code they wrote wasn't worth the effort of trying to understand the spaghetti inside. At the same time, some of them were disappointed when after the great effort they had seen, none of their code made it in. And some other guys just pass by, hear about this for the first time, and off-hand send Linus 30 lines of C, making Linus yell out of excitement. It's so wrong, ain't it?

Already at that time there were rumors that Linus accepts code more from 'cool guys' than from his own friends.

So by this time, he's thoroughly fed-up with that kind of crap. What could he possibly say to make folks believe it's the code itself, and its merits and nothing else?

> But Linus has both the right to choose or reject whomever
> he wants as a co-developer, and the right to say "no" to anyone he
> chooses.  Upon reflection, it seems like Linus has simply constructed
> a situation that allows him to avoid confrontation and to make him
> feel like he's doing something really special.

He's not one of those ego-guys. (Ego-guys in my book being Larry Ellison of Oracle, Richard Stallman of FSF, etc. Exemplary non-ego-guys being (surprise!) Bill Gates of Microsoft, Steve Bourne of unix sh fame, etc.)

So "to make him feel like he's doing something special" doesn't apply to Linus. Then again "constructed a situation that allows him to avoid confrontation" definitely applies to Linus.

By this time, he's not even trying to fend off accusations of egotism, nepotism, you name it. It's a lot easier to proclaim wide and loud that he's an asshole, an impossible being, and a capricious dictator.

That way he gets at least some room to breathe.

Besides, when did one ever see a real asshole doing anything else than cover up _being_ an asshole? Let alone "admitting" it publicly.

> He's welcome to do
> that, but I hope people realize what's actually going on when his bug
> fixes take five times longer than they need to.

(Wish I had something to say about that, but I don't.)

>> The generalization occured as a result of Georg Wrede's post in
>> which he took Linus opinion and expanded it into a distaste for
>> debuggers in general.

Aww, I was trying to make it not come out like that. I haven't anything against debuggers, it's just that I personally can't find them useful enough given my particular "brain anatomy". And I do see why it's important that kernel patches get submitted by Real Programmers only -- or we'll have to wait for bug fixes to take even longer.

(Suppose D gets as well known as Linux. Then it might happen that Walter's postman has to get a big truck for the fan mail and the Friendly and Helpful Patches. And he'll get some 2000 patches every day by email, not counting fan mail, folks asking for favors, and the usual crap.)

> Aye.  "Real men don't use debuggers."  Personally, I can't relate to
> Georg's comment about how not using a debugger forces you to think
> about the problem in a different way.  But then I don't use the bulk
> of the fancier features provided by debuggers these days either, so
> perhaps that's part of the reason.  And with the state of things in
> D, I haven't used a debugger here at all :-p

Heh, so D is making us all into Real Men! I kind of like that! :-)

John Reimer wrote:
> In article <43FFBB0A.9080007@nospam.org>, Georg Wrede says...
> 
> 
>> Heh, I read the article, and found myself agreeing with every point
>> in it.
>> 
>> Kernel work is such a serious thing that the faint should be gotten
>> rid of as fast as possible. And the bit about not having a debugger
>> forces you to think what the _program_ does vs. what a single line
>> of code does, was just succinctly said.
>> 
>> At the end I found it signed Linus. I was proud to have agreed on
>> the points, and I was ashamed to not having noticed the author's
>> name right at the top.
>> 
>> ---
>> 
>> Now, D is IMHO an excellent First Language. And some people really
>> seem to find debuggers useful. Personally I never use them, because
>> I find myself distracted and concentrating on the debugger instead
>> of thinking of the problem.
>> 
>> With D, let folks have debuggers if they want. But to become a real
>> programmer, one should learn to bike without the baby wheels.
>> 
>> Then again, not everybody should even try to become a Real
>> Programmer. Even the VB crowd can find jobs suitable for them. In
>> the future, also in D.
> 
> You mention a lot of pride here.  Are you from Finland, by any
> chance, Georg? ;)
> 
> I see you share more than one thing in common with Linus to be proud
> of. :)
> 
> Otherwise, I'm sure your opinion is worth its weight... besides the
> ticklish jab at what consitutes a "Real Programmer."

Crap! You've exposed me! Now I can't say anything positive about Linus without everyone getting suspicious!

Georg Wrede wrote:
> Sean Kelly wrote:
> 
> Problem was, from day one, that these guys wanted so bad to help, but the code they wrote wasn't worth the effort of trying to understand the spaghetti inside. At the same time, some of them were disappointed when after the great effort they had seen, none of their code made it in. And some other guys just pass by, hear about this for the first time, and off-hand send Linus 30 lines of C, making Linus yell out of excitement. It's so wrong, ain't it?

It's definitely a tough situation to be in.  I've thought about this situation even with respect to Ares (mostly speculatively, given that it's a tiny project) and I don't know that there's any way to make everyone happy.  So I can definitely sympathize with Linus if this was a motivating factor.

>> But Linus has both the right to choose or reject whomever
>> he wants as a co-developer, and the right to say "no" to anyone he
>> chooses.  Upon reflection, it seems like Linus has simply constructed
>> a situation that allows him to avoid confrontation and to make him
>> feel like he's doing something really special.
> 
> He's not one of those ego-guys. (Ego-guys in my book being Larry Ellison of Oracle, Richard Stallman of FSF, etc. Exemplary non-ego-guys being (surprise!) Bill Gates of Microsoft, Steve Bourne of unix sh fame, etc.)
> 
> So "to make him feel like he's doing something special" doesn't apply to Linus.

I didn't really mean it in that way so much as that I had a vague impression that Linus is somewhat insecure in the way that a lot of geeks are.  But I know hardly anything about the man so it was a shot in the dark :-)

>> He's welcome to do
>> that, but I hope people realize what's actually going on when his bug
>> fixes take five times longer than they need to.
> 
> (Wish I had something to say about that, but I don't.)

And I might be completely off the mark.  When you're as familiar with the code as Linus probably is with the Linux kernel, debuggers are often simply unnecessary.  But they're a great tool to have available just in case.

> Heh, so D is making us all into Real Men! I kind of like that! :-)

"The D Programming Language - Puts Hair on Your Chest" might be a bit of a put-off for the women here ;-)


Sean

Georg Wrede wrote:
> John Reimer wrote:
>>
>> You mention a lot of pride here.  Are you from Finland, by any
>> chance, Georg? ;)
>>
>> I see you share more than one thing in common with Linus to be proud
>> of. :)
>>
>> Otherwise, I'm sure your opinion is worth its weight... besides the
>> ticklish jab at what consitutes a "Real Programmer."
> 
> Crap! You've exposed me! Now I can't say anything positive about Linus without everyone getting suspicious!


Ha... yes, I suppose that was rather surreptitious of me. :)

I'm glad you forgive me. ;)

-JJR