Jim Starkey wrote:
> Jay Pipes wrote:
>> Jim Starkey wrote:
>>> Jobin Augustine wrote:
>>>> it appears that you hard core hackers never uses IDEs.
>>>> anybody using them ?
>>>>
>>>
>>> It all depends on what you call hard core hackers.
>>>
>>> Originally, hard core hackers were the guys that trampled the cutting
>>> edge of technology with audacity.  They pushed the limits and weren't
>>> afraid to experiment.  They developed OSes like Tenex and Unix,
>>> eschewing assembler in favor of high level languages, even when they
>>> had to invent the language to so.  These hard core hackers were
>>> willing to try anything that appeared on the scene -- objects,
>>> exception handling, formal interfaces, and, yes, IDEs.  The the stuff
>>> that worked got used and the stuff that didn't got left behind
>>> (Objective C, anyone?  Ada?).
>>>
>>> If you mean the contemporary definition of hard core hacker --
>>> arrested development adolescents -- then no, they're content to use
>>> the first rock they picked up to bang on other rocks, and are content
>>> to use vi and line mode debuggers and are perfectly happy to reset
>>> the same damn breakpoints after every build.  They also think that an
>>> 800 lines of nested flag testing is the very pinnacle of software
>>> engineering.  These are the guys who can take three years and
>>> hundreds of man years to put out a maintenance release with more bugs
>>> than when it started without asking what went wrong...
>>>
>>> I can't fathom why some troglodytes are so proud of the ignorance of
>>> Windows that they're unwilling to even try a different development
>>> technology.
>>>
>>> Oh, well.  At least it keeps these people from creating new products.
>>
>> I sincerely hope you weren't referring to me or anyone in the Drizzle
>> contributor community in the above statements.  We try on this mailing
>> list to keep the humour above the belt.
>>
>> -jay
>>
>> p.s. For the record, I was a software developer using Windows
>> platforms for years before I began using and developing open source
>> software. And, yes, I used VisualStudio and other IDEs when I did so. 
>> That said, I am more productive and efficient, not to mention more
>> knowledgeable about the underpinnings of software, now that I do not
>> use an IDE and instead use the tools of the GNU toolchain and vim. 
>> But, this is my own personal opinion of my productivity.  Everyone
>> makes there own choices about what tools make them most productive,
>> and there's nothing wrong with that at all.
>>
> No, I wasn't referring to drizzle, I was referring to MySQL.  But still,
> drizzle remains, by choice, in the dark ages.  I continue to be
> astonished that people care more about line length than reducing the
> amount of spaghetti code by encapsulating things like values and
> exceptions and recognizing that "plug-ins" that have to be not only
> compiled on the code base, but compiled with the code base are worthy of
> the name.  

I'm not sure it's fair to say we care _more_ about line length. We spend
enormous amounts of time reducing spaghetti code by encapsulating
things. I've been actively working on out-of-tree plugins and should
have the last little bit of that done this week in fact.

We do care, however, about line length and other code-style issues,
because when there is more than one developer working on a codebase,
some of those things need to be actually discussed out loud. (you know -
because none of us intrinsically agree and the decisions are arbitrary
anyway, so they need to be made out loud)

> Drizzle has been slightly successful in reducing complexity
> by shedding badly designed components, but has been seemingly unwilling
> to take on the more important issues of modularization, encapsulation,
> streamlining of execution phase, etc.

I'm not sure which Drizzle you're talking about here, but it's obviously
not the code base I'm working on. Not only are we not unwilling to work
on that, a good amount of active progress has been made and continues to
be made. You may remember that spaghetti code you mentioned earlier...

> Over the summer I came to the conclusion the NimbusDB development would
> never be predictable without a multi-process / multi-node / multi-thread
> point and click debugger. 

Neat. Is it published anywhere? I'd be happy to package that up for
inclusion in Debian to help push forward the state of the art.

> So I wrote one.  Even though the target
> environment is 64 bit linux, it was vastly faster and easier to write
> and debug the GUI, elf, and dwarf components in Visual Studio.  The
> Linux elf standard turns out to be a pack of lies, so I had to spend a
> great amount of time running gdb compiled for debugging under gdb to
> scope out reality (hint: the frame description entry (FDE) in the ELF
> unwind section has an undocumented subfield), so I'm a little sensitive
> on the subject of debuggers and the relative productivity thereof.
> 
> 


_______________________________________________
Mailing list: https://launchpad.net/~drizzle-discuss
Post to     : [email protected]
Unsubscribe : https://launchpad.net/~drizzle-discuss
More help   : https://help.launchpad.net/ListHelp

Reply via email to