discuss-gnustep
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

The evidence against UIKit (Re: Kickstarter was not successful... but it


From: Doc O'Leary
Subject: The evidence against UIKit (Re: Kickstarter was not successful... but it did help things...)
Date: Thu, 26 Dec 2013 16:23:04 -0600
User-agent: MT-NewsWatcher/3.5.2 (Intel Mac OS X)

In article <mailman.10332.1387921990.10748.discuss-gnustep@gnu.org>,
 Tom Bruno <thomas.bruno@me.com> wrote:

> Science research, especially when conveyed to the target audience requires 
> more than just "because I say there is evidence".  You go on about how its 
> science yet I've read this thread over and over and I can't pick out a single 
> reference, scientific method, or any actual research from you. It is "because 
> I say so".

*Sigh*.  So even responding with evidence gets this kind of reaction?  
Absolutely *nothing* I have offered is "because I say so"; mine has only 
been a direct contradiction of the GNUstep *authorities* statements of 
"because *we* say so" non-scientific approach.  I can't help you if you 
refuse to not only listen to me, but refuse to accept the evidence the 
world is screaming at you.

> I've read through the website and I don't have any question about the 
> direction of GNUstep. What I'm hearing from you is that you don't want that 
> to be the direction.

Then you need to listen.  Outside one side of GNUstep's mouth comes the 
message that you can use it to port Mac apps.  From the other side of 
the mouth, Mac support is listed going from "use MacPorts" to 
"unsupported" to being completely unmentioned!  And that's just *one* of 
the mixed messages.

> I disagree completely, as my intent for using NS* and 
> UI* in GNUstep matches the stated goals completely.

Then you'll have to help me out by supplying a link that covers UIKit 
*at all* on gnustep.org.  I certainly don't see it on the main page or a 
couple levels deep.  Since you seem to find the direction oh-so-clear, 
clear it up for me, too.

> I don't care that you 
> want UIkit to wrap NS*, I don't care that you think UI* and NS* are the same. 

I have never said that.  Building straw men undermines your position.

>  They are not the same and anyone with a solid background in architecture 
> would understand there is more to the issue than API calls. Since you have 
> not demonstrated any experience in architecture and propose a model of how 
> things work that is contrary to my first person experience, Please forgive me 
> if I don't take you at your word.

I have never asked you to.  All I've done is state some question as to 
what *should* be the architecture of UIKit support.  If you really care, 
here is an outline of my thinking.

Now, for Apple, it probably made a lot of sense.  They have all sorts of 
research efforts going on all the time, and most of them shouldn't begin 
with the idea that radical change *must* fit into an existing framework.  
Prior to the iPhone taking off, it wasn't even a given that either 
mobile or touch would turn out to be a successful direction.

So what it *seems* like the starting idea was is to take Cocoa and fork 
it, doing whatever it takes to develop a good touch/mobile experience.  
That sort of evolution would explain a lot of the "surface" differences, 
like the switch to the UI- prefix, or the table views that, in the end, 
behaved so much *unlike* a table that they would have been called 
something different if they had been designed ahead of time.

But it's different for GNUstep now.  We're in a position to take Apple's 
hard work and fit it into place after the face.  And not just from the 
perspective of Cocoa, but in the larger world of overwhelming success 
that both mobile and touch technology has experienced over just the last 
5 years.

On one hand, we can see that, despite some tweaks, touch itself doesn't 
change much.  It is just a new set of gesture events that map to some 
underlying behavior (e.g. pinch to zoom).  Indeed, an iOS developer 
running an app in the simulator can "touch" quite readily with the 
mouse.  On on the flip side, multitouch track pad are available for 
desktop machines, leading people to expect all kinds of apps to allow 
gestures.  Even the underlying behavior can be triggered by now-common 
hardware, such as using control-scrollwheel to zoom.

>From the "touch" aspect, there seems more to be gained by a unified 
architecture than a split one.  For the "mobile" aspect, it is less 
clear to me.  Is it a goal of the GNUstep project to support devices 
with very small screens and other major resource limitations?  If so, 
then there is indeed a lot more work that would have to be done to fit 
into the interface paradigm that is imposed by that kind of hardware.  
Perhaps even enough to warrant a wholly split API.

So if you actually want to make a case rather than just beating your 
chest with nothing more than assurances of "first person experience" 
while ironically pretending it is *me* that offers "because I say so", 
that is where you should start.

The rest of this really has little to do with UIKit at all, so people 
can stop reading here if they don't want to see me rehashing old points.

> As stated above, pretty much everyone reading i.m.o. can't figure out what 
> this evidence is. You've made a few good points, but even those points aren't 
> backed by scientific research and are just repeated. If they are based on 
> science then you are not demonstrating the methods to arrive at the 
> conclusions to make your case believable. (A requirement of Science)

One of my points continues to remains that many here have no interest 
whatsoever in looking at *any* kind of evidence.  Their entire 
methodology is wrong.  They, like you, *start* with the conclusion and 
then go looking for things that support it and ignoring things that 
refute it.  I start with the expectation that *I am wrong* and go about 
gathering evidence of my own, both formal studies and anecdotal, that 
allows me to either *confirm* that (in which case I change my thinking), 
or *refutes* it, allowing me to continue to be a blowhard.  ;-)

Either way, the burden to get *your* house in order still depends on 
*you*.  Don't take my word on anything, and that includes the methods I 
used to arrive at my conclusions.  Look at the evidence, *all* of it, 
for yourself and tell me if your conclusion differs.  If it does, then 
and only then is a discussion of methodology going to be meaningful.

> Most Kickstarters that are a success are a success because they 
> have already earned an audience of individuals before the kickstarter. This 
> pre-populated seed group are actually the ones that end up funding the 
> project, viral kickstarters are the edge case.

You support my point.  GNUstep has failed to engage a seed group with a 
meaningful message, making it next to impossible to communicate that 
message to a larger audience.  Why are you fighting with me when you 
substantially agree?

> Using kickstarter as evidence that any specific characteristic of the project 
> is the cause of kickstarter failure, is a failure to analyze and present the 
> data in a fair and unbiased view. Infact, one could call the Kickstarter a 
> huge success since it brought awareness and myself to GNUstep.  Not all 
> kickstarter goals have to be for $.

I never said they did.  I only ask that you provide *evidence* of 
success, because it sure looks like a failure to me.  Other than some BS 
about "awareness", what metrics of engagement have been measured?  Was 
there a *massive* influx of new developers, or just a handful or just 
you?  Have there been *massive* organizational changes, or is it still 
running the same way?

> Actual References:

http://www.google.com/trends/explore#q=GNUstep&cmpt=q

Ouch!  Instead of simply proclaiming success "because I say so", tell me 
how you meaningfully measure it.  Because, right now, your word isn't 
worth very much to me.

> Objective-C itself has not really changed all that much in the 
> last 5 years. In fact I was very surprised when I got GNUstep up and running 
> just how complete it is!  ARC, Blocks, a basis in CoreAnimation, this is way 
> more than I expected!  Fantastic work guys.

I had the same feeling back in 1996.  And that is why it's always been a 
bit sad that nobody outside the group cares, and nobody inside the group 
seems to care that nobody outside the group cares.  The problem clearly 
have little to do with code, but much to do with coders who don't seem 
realize that sometimes you have to work on more than code.

> You want to see not very welcoming 
> then go experience the Gnome project for a bit.

It is never a good defense to say "At least we're not as bad as *that* 
guy!"

> OSS require a small layer of 
> thick skin and for what its worth I am very proud of how Gregory and others 
> have treated you throughout this whole thread.  They've shown extreme 
> restraint & that itself is evidence that this community is not only welcoming 
> but very mature and open.

Thank you for that chuckle.

> You've wasted so many peoples time and have proven nothing about your skills, 
> your research, your hypothesis, or your scientific method.  If you hadn't 
> noticed, this is a software project not a religion.

Re-read those two sentences.  I have, indeed, tried to *not* make this 
about me, but rather about how the approach of GNUstep is fundamentally 
flawed.  If not flawed to the point of religion, then at least to the 
point of being a cult of personality.  If you wish to reject the 
arguments I give solely on the basis that I'm not willing to participate 
in that cult, then that continues to say more about you than it does 
about me.

> As far 
> as your thoughts about "leadership" I disagree respectfully as my experience 
> shows that Management can say anything, but the guy that knows how to code is 
> the one that all the coders follow to get the job done.

And in my experience, a lot of projects turn into death marches because 
the coders don't have the guts to push back when management is wrong.  
Or the foolishness of those to whom "get the job done" means a poorly 
thought out mess of spaghetti code.  Point being, simply being able to 
write a computer program doesn't give you any special authority either.

> So what, you don't agree with leadership. Take control of your own life and 
> destiny. Make it better. Just do it.

And thus the "it" I continue to "do" is "just" something that barely 
intersects with GNUstep.  It should not be that way.

> Your arguments are un-motivating, 
> uninteresting, and are soaking up time

Then simply tell me to stop if you see no benefit in discussing this.  
It will happen in the due course of time if nobody is interested in 
changing their approach.  One word from you and I'll go away *now*, and 
you'll be known as the hero who vanquished me.  Go ahead . . . just do 
it.

>  There is no profession where you get to say "I'm awesome" without proving 
>  it.

And this is the funny backwards world you're in.  It is *you* who is 
saying that!  I'm just the guy who's looking on from the sidelines 
saying "You don't look so awesome to me."  And, oh, the tantrums that 
have been thrown because of this!

> Why you think we'd all just take you at your word is beyond me. You 
>  remind me of "that guy" who sits in every meeting and hijacks every code 
>  strategy talk so he can say later that he contributed even though he didn't 
>  write a line of code.

Do you know what the word "strategy" means?  You do realize, right, that 
an umpire doesn't need to be a pitcher *or* a batter to call a strike?  
But if you want to start a pissing match, you go first: what code in 
your body of work can you show me that will make me defer to your 
expertise?

> The coders would have succeeded at writing it a whole 
>  lot faster if "that guy" didn't keep extending the discussions to never 
>  ending points holding up progress. Perhaps some self reflection is in order 
>  after reading Dilbert?

Do whatever you need to do.  I self-reflect all the time.  Sometime I 
come to the conclusion that "progress" means more than just going 
further along the wrong path.  So see, my default approach when I'm 
thinking is to assume that I am *wrong*.  What's your approach?

-- 
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.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]