Re: ::scr open sores

[Simon Batistoni <simon@xxxxxxxxxxxxx>]

On 14/01/02 14:56 +0000, simon wistow wrote:
> On Mon, Jan 14, 2002 at 01:21:49PM +0000, Simon Batistoni said:
> > Gah, quit moaning you fule. If you really want the moon on a stick
> > then I'll place an order with NASA, but you could try just calming
> > down a bit first :) 
> 
> It may have come across as moaning but, the point i was making was that
> OS software is not the mutts nuts that it's been hyped up to be.

*Ah*. You're fed up with the blind nathanly worship of all things
with Linux or OSS written on them when, in fact, Linux and OSS is
just another load of software, and All Software Sucks.

Well, yes, obviously, but buying hype about *anything* will get you
disappointed. Starbucks make rank coffee, the next Radiohead album
will probably be Pink Floyd prog-rock arse, and Gap clothes are all
basically horrible.

> > Different interfaces suit different people. 
> 
> True. But there are metrics for suckiness. Most OS software fills that
> because they don't have professional interface designers. If there was
> no need for them then companies wouldn't employ them. (There are jokes
> here about lawyers / marketers etc but I'll let 'em go past)>
 
> Have a look at 
> 
> http://developer.gnome.org/projects/gup/ut1_report/report_main.html
> 
> for OS interface suckiness. 

Yes, although I find that quite heartening at the same time. I only
really have time to skim it, but it seems to show the Gnome team
getting the Sun HCI team to look at Gnome and run some user studies.

> > The human mind is simply not capable, whether working
> > singly or as a team, of creating a nicely laid-out, coherent complex
> > system. 
> 
> I'd disagree. But I have no proof. I cannot scream.

Well, I'd like to think differently, but I have never actually seen
it done. All software systems seem to go the same way. Just look at
the broad sweep of examples you gave in your thread-starter, across
various types of program and system.

> > Things need to work quickly, to meet business or other deadlines.
> > Quality gets sacrificed for speed. This will always happen.
> 
> I was actually going to use this argument for Windows GUI vs *nix GUI -
> the DRI is in the kernel which means speed but dodgy graphics drivers
> can theoretically take down the kernel. 

Oh, well I was talking more about speed of development time. As in
"we need to get to market next week, so finish it *now*".

> As for one bad experience, I'm not trying to cock wave, but it's not
> true. I've never like any of the *nix WMs that I've used from CDE, FVWM
> (and FVWM 95) through Afterstep, Window maker, Enlightenment DR14 (which
> I wrote patches for) and all subsequent versions (DR17 looked like it
> could rock but ... well, see that last URL I posted in my second mail) 
> up to Gnome (which I run on my machine at home).

Fine. But the most specific I've ever heard you be about why Linux
sucks is that the Mandrake install you did at work was a nightmare.
What are the main things that you find sucky about all of the above
(or, perhaps more easily, what are the things that are missing), and
how would you like to replace them?

> > But here's where open source is great. You can quit moaning, and fix
> > what you think's wrong. 
> 
> I wish. Yes, theoretically I could patch bits here and there. Especially
> if they were small bits. I patched bugs in E and added functionality so
> that there was an option you could set that would make double clicking
> on the window bar iconise the window rather than window shading it. But
> patch the whole WM? Is that honestly realistic? That point is often over
> looked by the OS community. 

Well, no, because you get to the point where you may as well write
your own WM that works just the way you want it. I just think that
being a little more constructive - trying to work out what's wrong,
and how it might be fixed, would be less frustrating than bashing
your head against the "software sucks" wall ad infinitum. 

I'm not saying you can patch the entire WM single-handedly, but you
can at least *try* to make a difference. 

I agree with you that the OS movement massively over-hypes things,
but over-hyping doesn't actually mean that there are good points to
the core idea.

> > > * GCC
> > So keep the good bits and rebuild the rest.
> 
> ??? That doesn't make sense. Which bit would I keep? The Front End -
> Back End decoupling?  Then I'd have no parser and no code generator. 
>
> 
> Anyway, true, I could write my own compiler, but see above for reasons
> why, in actuality, I couldn't. 
> 
> It may be possible, it's not practical.

Well, maybe not. But perhaps you could write to the GCC mailing list
(like what they ask you to at http://gcc.gnu.org/about.html) and
raise the issues you have there. Yes, you might be shouted down by
15 year olds who think Stallman is God, or lost in a general fuzz of
noise-to-signal. But there are feedback mechanisms in place, and
there is at least a chance to have your say.

> Anyway, my point was (in a sprwaly, rambly way) that OS isn't all
> it's cracked up to be. Despite Eric Raymond's protestations that it
> would revolutionise software engineering I think it's failed to do
> that.

Raymond and Stallman are both cultists at heart, and the trouble
with any blindly-followed cult is that it gets offered as the "magic
bullet" for everything, and is defended to the death by its
adherents. Whether it's OO, XP, Church of the 9th Day Nutters or
anything else.  We've had this conversation before, I'm sure...

sb