Re: Kickstarter was not successful... but it did help things...

[Doc O'Leary]

In article <mailman.10183.1387826241.10748.discuss-gnustep@gnu.org>,
 Ivan Vuãica <ivan@vucica.net> wrote:

> This isn't criticism, this is demands.

Such is the nature of science.  There is a basic rigor that must be met.  
This remains true for computer science.  Do you cry like a baby when a 
compiler demands correctness?  Is throwing exceptions an example of "be 
positive"?  In science, criticism often rises to a level above that of 
mere opinion.  If you don't like that, you're in the wrong field.

> This will be my final response to you in context of this discussion. I
> would invite others to cease responding as well, as your criticism is not
> of the constructive kind.

On the contrary, it *should* be the *most* constructive kind.  Look at 
it this way:  freedom of speech isn't needed to protect comfortable 
ideas.  You don't face much risk in saying "Cats are fuzzy!"  What you 
need protection for is when it is necessary to say of the Emperor "But 
he isn't wearing anything at all!"

Ignore me an I *will* fade back into the woodwork.  If you think that is 
a good thing, if you think that *fixes* the problem, you have a lot of 
growing up to do.


> Your criticism serves itself, as you refuse to believe that a direction for
> GNUstep cannot be set, that everyone has their own goals, and that people
> will not submit to a central will. I won't suddenly work on themes just
> because the project direction is "we need better themes" (which we are
> luckily getting).
> 
> I'd say "think about it for a while", but you will claim that you have, and
> that the entire development team is unreasonable for not bowing and
> submitting to a central development plan, to a central authority, to a
> single Great Leader(tm) who shall show us the path to
> glory<http://kimjongillookingatthings.tumblr.com/>

This is all a straw man.  I have never called for any of this.  Keep 
your fantasies inside your head.

> Apple's reasons includes improving consistency of the APIs and
> reorganization in the world of constrained devices that nonetheless run
> hardware-accelerated graphics. If you use UIView the same way you use
> NSView, as you claim, you might be doing it wrong; but if you plan to use
> UITableView the same way you use NSTableView, you're certainly doing it
> wrong. With UIView, you should plan for it being backed by CALayer and
> hence a texture that is expensive to upload, but cheap to paint over and
> over again. With modern NSView, you should do the same, but the API itself
> does not fully count on that (especially back in 2007). UIView does.

You're just not getting my point.  I'm talking about NS- vs. UI-, and 
you're talk about all manner of other other things that are either 
deeper than the API or unrelated.  Yes, Apple has made things like table 
views behave radically different on both platforms, but it isn't the UI- 
prefix that matters (i.e., they could have called it NSListView).  And 
from a portability stand point, of *course* you're not just going to do 
a s/UI/NS.

> Apple's reasons include forcing developers to think in a different way.

Oh, nonsense.  A better argument is that they simply wanted to get away 
from the NS- prefix because, as I've already covered regarding GNUstep, 
the reference is so dated that it is meaningless.  I'd wager that the 
majority of iOS developers have absolutely no clue why the Foundation 
classes start with NS-.  They're not thinking about it *at all*, so 
there is no need to push them to think different.

> Apple has a history of manipulating developer behavior by deprecating older
> APIs, introducing new ones, and using architecture switch to completely
> remove the old behavior and simplify their own codebase. See: the removal
> of BlueBox upon switch to Intel, removal of PowerPC emulation support in
> 10.7. UIKit is just another way to manipulate developers' behavior while
> simultaneously getting rid of cruft.

Quite possible, but not what I would call a *good* reason.  This is the 
kind of thing Apple does that we *don't* necessarily want to duplicate.  
In the larger scope of things, though, you have to put plans in place to 
deal with these realities.

> Do you need more reasons for why Apple took the UIKit route?

Yes, because I still haven't seen any good ones, which is what I asked 
for.  More to the points you're making, you have to look to the future 
of GNUstep.  If Apple is possibly prepping for an overhaul or 
deprecation of AppKit, then you're *really* in a pickle.  The UIKit 
direction is something that should have been addressed by GNUstep 5 
years ago.

> For me, a simple reason for going the UIKit direction is that today there
> is a way to run (some) OS X applications on free systems, but there is no
> way for way more numerous iOS apps to be ported to free systems (and
> beyond). Commercial offerings exist, and they will always have a reason to
> exist and provide added value. But let's have a free one as well, to help
> students, to help free software users, et cetera.

Of course.  As you have come tired of hearing me say, the main problem 
is that nobody steering the GNUstep project seems willing to plant that 
flag.  Such an omission certainly didn't help the Kickstarter campaign 
any.

> Pick a direction for yourself and walk in that direction.

Do you imagine that I'm incapable of that?  What do you think I've been 
doing since 1996?

> Don't expect us
> to tell you where to go.

The direction I will go is "away from this".  It's what most developers 
already do; they don't even care enough to criticize GNUstep.  You can 
bitch about my being here bitching, but that is the reality you're faced 
with.  Inbreeding is a dead end.

> Taking no action is not an action.

And, indeed, the driver of a car does not *drive* it.  They pilot the 
vehicle, which the *engine* drives.  I *drive* a bike, but nobody calls 
it driving.  They call it "riding".  Just like a passenger in a car.  
Funny, isn't it?

Rest assured, that I fully understand that the vast majority of people 
involved with GNUstep, just as with most open source projects, fail to 
see anything other than code as action.  I keep looking for that to 
change.  Not yet, I guess.  Seems like I'll have to keep riding.

-- 
iPhone apps that matter:    http://appstore.subsume.com/
My personal UDP list: 127.0.0.1, localhost, googlegroups.com, theremailer.net,
    and probably your server, too.