Re: Feasibility/Requirements/Wishes of xAL (was: Standard Library Interest?)

[Marin David Condic <nobody@noplace.com>]

Stephane Richard wrote:
> 
> *** As far as the Ada Community, vendors, and others.  I can't really say I
> haven't had much interactions with the related businesses.  But in general,
> in the software industry, I do have to agree with you.  it's all about "give
> me what I want, even if I do know what I want yet" Impress me then we'll
> talk.  This almost forces other companies and developers to program in their
> basements, try to revolutionize the world and see if it works, lots of
> wasted time in so many cases that I can understand your point of view on
> this.  looking at the writers guild however, it's the same thing, people
> spend alot of time wrinting, in hopes that a producuer will want to make a
> film out of their stories, get a best seller, etc etc....There is a reason
> why the industry has gotten how it is, an dI was right there to see it
> happen too...And although I can say you're right, I find it hard to see a
> reason why we shouldn't try to revolutionize the world.  My Motto is if
> that's what they want, than that's what they're gonna get basically.  Sure
> it's a chance, the question is how big a change is it?  even outself the
> vendors, if we have organizations to which the vendors respect working on
> it, or approving it, our chances of influencing the vendors jsut got
> multiplied by many folds.  it's a two way street.  We either influence the
> vendors to influence the WG9, or we influence the WG9 to influence the
> vendors :-).

You're making it too complicated. The vendors themselves undoubtedly 
have some ideas about what they'd want to see in a library already. You 
just need some "general" notions - not detailed requirements. You'd get 
the same kind of thing by surveying some subset of their customers. 
You're not trying to get *perfect* alignment - you're just trying to be 
sure you are working in the right general direction. You'll close the 
loop and keep honing the end product along the way.


> 
> *** I couldn't agree with you more, ada HAS had enough of that and needs no
> more.  How do we solve this?  Sure there's that secondary reason, but that's
> just one avenue, like you mentionned elsewhere, there's the vendors and
> their customers, do we need to go through the vendors to get the customer's
> opinions?  Maybe not :-).  and we'd still get the results we need, make what
> the customers want.  The Market Research that the vendors might have done
> probably wouldn't answer all our questions anyway, a market research is
> usually very specific and asking "do you want or see yourself using a
> library" such as is proposed by CAL might be to vague a question for any
> research to arrive at an answer.  Library specific?  now you're talking.
> 
You don't need that much detail. You just want to be sure that if you're 
charging off to build a matrix arithmetic package that everyone wasn't 
screaming they wanted a container library *first*. So if you had a 
priority to build a library for capability X, you might wave some 
general ideas around in front of the vendors or some group of customers 
that would act as a beta test site and see what they say. You come back 
with some package specs or more detailed design for capability X and ask 
them "Is this close to what you had in mind? You just need to get some 
end-user involvement and feedback to be sure that you're not totally 
wasting your time.

> 
> *** to some point I do agree with that.  Well let's say I do agree that it
> takes motivation to reach the statue such as is aimed by CAL.  People aren't
> used to giving themselves a deadline when there's no paychecks at the end of
> it or some form of recognition for the work done.  How about if I give you a
> free Palm PDA? ;-).....in a lot of my personal projects I did manage to give
> myself a deadline and reach it, but I do have to admit that it's not the
> same context.  Let's say there was no one paid, at least not until
> everythign as matured considerably (read "prove your .... blah blah blah ;-)
> How could we motivate?
> 

It isn't just a matter of motivation. Its also a matter of being able to 
insist that certain deliverables show up and meet certain standards. You 
can always refuse to accept some work done by a volunteer because it 
doesn't meet standards, but you don't have much leverage to hold over 
the volunteer. If you go to the volunteer and say "I have this paycheck 
and if you want me to give it to you, you'll have to deliver X and Y and 
Z." (Even that doesn't always work! :-)

On one end of the spectrum, you have an all-volunteer army of developers 
and will likely get slow, chaotic results - but they will at least be 
"cheap" results. On the other end of the spectrum, you could hire a 
group of developers or outsoruce company to develop the whole thing by 
some given date and to some given standards. They don't get payments 
until the deliverables show up and pass inspection. You get really good 
control of the end result that way and you get it to meet your schedule, 
but you're going to pay through the nose to get it done. Between those 
two poles, you've got a variety of alternatives that will trade $$$ for 
relative levels of control. Get creative and try to figure out a way to 
do it with limited budget and you might have something there.

> 
> *** I'll let Robert answer this one ;-).  or anyone else in any related
> committee.
> 
My experience with committees is what I based my rectally extracted 
estimate on. I'm not criticizing committees, but I'm being realistic 
about what they can do. Nobody will have this as their full-time job, so 
it is at best a hobby, getting attention if and when time permits. There 
will be a dozen or more people involved, each with their own ideas and 
agendas and nobody will be able to say "I'm sorry all your great ideas 
and agendas are going into the trash heap, but *THIS* is what we're 
going to build..." (Since they're all volunteers, you've got to get them 
all to "volunteer" to go along with a given plan. At *best* that is very 
time consuming and at *worst* it never happens.) Realistically, the 
committee is likely to be separated geographically which makes having a 
"brainstorm" session difficult and makes it hard to get everyone acting 
as a team. Maybe yuou can get together for a few days once a quarter - 
if everyone on the committee is sponsored by their company or is 
independently wealthy enough to take the time off and buy their own 
plane tickets to support what we have already said is a "Hobby".

You can see where I'm going here. A committee under the aegis of SIGAda 
may be a useful thing for some parts of this, but it will move real 
*ssssslllllooooowwwwww* - if at all, and it could be *years* before they 
get any real results. They might properly serve as the "gatekeepers" - 
whoever develops the library has to wave it past them to get final 
approval to make it "Official". I just wouldn't see such a committee as 
being the source of production.

I think there are better ways and mostly they involve some form of 
collective development on the part of the vendors. Let *them* put 
together a team in some manner and fund its development in some way. 
They get the end product and can go put it in front of some SIGAda 
committee to get its acceptance as something with the stamp of "Ada 
Approved" on it.

> 
> *** Agreed, time is of the essence in this project.  everything depends on
> when we get something started.  If we can't wait 2 years (or more) for
> SIGAda to shake it's booty, can we also wait for vendors to become involved?
> or should we "prove to them" there's that word again that they would have a
> good reason to get involved?  Basically see my previous paragraph about
> Influencing one to get to the other.   Which side is both quicker (probably)
> and easier to accomplish?  taking we haev Robert that is already
> volunterring, you think he believes in the potential of this library?  I
> think so :-).  So do I.
> 

You can start developing anything you want. I have stated - and am 
sticking to it - that I won't do anything unless the vendors are somehow 
involved. The reason? I've seen this before and I've been involved 
before and it goes nowhere really fast with lots of effort out the door 
and little to show for it. I've stated here before that if you and I (or 
any other team of developers) got together and developed the most 
wonderful library in the world on speculation that we could get it 
adopted, its still just another one of many sitting on some website and 
with no more likelihood that it will become "The Thing" than any of the 
other libraries out there. If anything, it starts out behind the line 
since other libraries already have *some* users. If that's where you'd 
want to spend your effort, better we start with something like the Booch 
Components or one of the other existing libraries and try to get the 
vendors to accept *that* as the basis. It hasn't happened. It could have 
happened. Its not going to happen. More effort trying to make it happen 
is IMHO wasted.


> 
> *** There are things that Vendors could tell us about yes I'm sure they are
> after all vendors which gives them a range of information that we as
> individuals may not have, as a business dont you think that if they were to
> share this "exclusive" information with us they wouldn't want to be
> compensated for it in the same way we'd wanna be compensated for CAL?
> Corporate information is corporate information... Wouldn't they consider
> that an investment in the project itself?  even if the reseearch is already
> done by now and requires no extra man hours to produce a document would
> could benifit from?
> 
You're over complicating it. You just want them to tell you what sort of 
things their customers are asking for. They're the "consumer" of the 
library - why wouldn't they want to help you build what their customers 
want?



> Now understand that I'm not debating your ideas here, but my experience in
> the industry tells me that those are definite plausible answers you're
> likely to hear from the vendors and as such "vendors that is" would
> righfully allow themselves to state it.   Like you, and all programmers, who
> wouldn't wanna get paid the big bucks for a project like this one?  I'd be
> lying ot say I dont want money for this kind of effort. :-).  This is a
> capitalist nation after all ;-)?  the question is when should we expect
> compensation?
> 
I need to pay the bills. I can spend only so much time being charitable 
to compiler vendors and then my wife will start insisting I get away 
from the front of the computer and get out and mow the lawn. I'm willing 
to give *some* charity to the vendors and the Ada community, but *some* 
effort is fairly limited and it won't be enough to produce a really 
comprehensive library. I doubt I'm really unique. Do the vendors want a 
library? Do they want it bad enough to fund its development in some 
manner? No? Then Ada won't get one. That's my prediction.


> *** I think I can clear this one quick and easy...if CAL would ahve been
> taken how easy, today, would it be to come up with a new 3 letter name?
> :-)....One that's representative of what we want this project to be?  If I
> thought of a library and see the name is available, I'd jump on it to make
> sure I dont have to think up a  new name for it one that might make less
> sense than the original name :-).  if I could reserve it, I would :-)
> 
Lets see.... There's 26 ways to fill the first position, 26 ways to fill 
the second.... My calculator says 17576 names. And, if needed, I can go 
to four characters. 456,976. Or we could toss in numbers. Or I could 
stretch it out into a really long name. I really am not worried that 
someone is going to steal away "CAL" from me and produce a competing 
library. I just don't know what we're going to see with some kind of 
registry that really matters that much at this point. I could be 
persuaded - but I don't understand what is being accomplished.

MDC

-- 
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jsf.mil/NSFrames.htm

Send Replies To: m c o n d i c @ a c m . o r g

     "All reformers, however strict their social conscience,
      live in houses just as big as they can pay for."

          --Logan Pearsall Smith
======================================================================