Re: libraries for Ada (was): periodicity

["Marin David Condic" <dont.bother.mcondic.auntie.spam@[acm.org>]

Boy, that sure is an ambitious list! :-)

I agree that some standard libraries would be an excellent addition to Ada
and would serve to make it more attractive. However, I'd make the scope a
little narrower in the hope of actually getting it adopted:

Some version of math libraries - including statistics & matrix/vector math.
(Insert your favorite math needs here, but I'd stick to the branches of math
that are most frequently used first. Stats & Linear A seem to be pretty
popular across the board, but there may be others as well.)

Some version of basic data structures. Could we get the GRACE name settled &
maybe see if a reference implementation can find some level of acceptance?

Some version of a GUI - I'd base it on XML & possibly third-party
interpreters. (I hear there's kind of a standards effort to do this at
http://www.mozilla.org/? Or am I mistaken?)

Shoot for that much & I'll bet we have a substantially better language.

MDC
--
Marin David Condic
Senior Software Engineer
Pace Micro Technology Americas    www.pacemicro.com
Enabling the digital revolution
e-Mail:    marin.condic@pacemicro.com


"Kent Paul Dolan" <xanthian@well.com> wrote in message
news:171516c6068daa0fd32844182995dd3b.48257@mygate.mailgate.org...
>
> I suspect each person's choices would be colored by their needs, and I
> _know_ I am going to regret getting into this, and I'm _sure_ I'll hear
> "but that already exists as (obscure, non-standard hobbiest) library X",
> but for starters:
>
> Ports of the IMSL math and statistics libraries.
>
> Libraries of accessor/mapping routines to the various popular kinds of
> close coupled, high bandwidth, low per processor capability massively
> parallel processors, and same-named adapters for doing the exact same
> problems where the underlying hardware is a network of slow
> communicating, high capability computers (like SETI at Home).
>
> Pretty much all of Java, most importantly in the Java style of being
> well integrated and astonishingly well documented, but compiled instead
> of byte-compiled, so that for example students can just sit down and
> build GUIed applications using the Ada language "right out of the box".
>
> A (heavy emphasis here, contrasting to the partial implementation of GKS
> some years ago in Ada) _full_ implementation of PHIGS or of a freely
> licensed RENDERMAN or some similar high tech graphics library, for
> photo-realistic animated rendering.  Similarly for animation.
>
> All of the algorithms published as the Graphics Gems series.
>
> A standard set of graph tools for trees, graphs, digraphs, et cetera,
> and all the usual operations on them.
>
> A knowledge engineering library with support for the standard knowledge
> capture paradigms and structures.
>
> High technology text and word processing component libraries.
>
> Data presentation libraries.
>
> Data mining and indexing component libraries.
>
> Evolutionary algorithms and neural network component libraries.  (I'd be
> more than happy to port my recent algorithm inventions here to Ada and
> contribute them.)
>
> Data base functionality tightly couple to the Ada language but
> supporting e.g. SQL standard protocols.
>
> A game-writing toolkit.  Arguably, in terms of gaining wider acceptance
> for Ada, this should be the top priority, as it would attract a crowd of
> students and individual freelance programmers, exactly the crowd that
> made C a success outside of its original laboratory environment.
>
> Financial management, planning, and estimating tool kits.  Stock market
> predicting components (might as well throw in snake oil, it keeps the
> snakes well oiled, why not the rest of us?)
>
> Network programming, Web search, and Web presentation technology
> libraries.
>
> Essentially, one could take all of the various ACM and IEEE computer
> oriented proceedings, start at volume 1 number 1 of each, and work
> forward, adding each published algorithm to a library, while
> regularizing data representations and such to make the libraries
> interoperable.
>
> Lots of other people would have longer, more practical lists.
>
> The points that get buried in list making are that there needs to be a
> start, that it needs to be organized, that it needs some hope of
> longevity via a sponsor that is more trustworthy than, for example, the
> US Military's defunct Ada Mandate, it needs a home from which it will be
> accessible, like the WUSTL archives, it needs librarians and indexers
> and master programmers to keep it well organized and easy to search for
> needed tools, it needs a proofing and filtering input stage where the
> community can make contributions and know that they won't just
> disappear.
>
> xanthian, "duck and cover" time.
>
> [Did I mention "free"?  It is pretty pointless to price oneself out of
> the market when trying to foment a revolution, as Sun does by putting
> license fees on commercial use of Java, essentially preventing the
> hobbiest from doing a slow transition to commercial success by putting a
> huge "first step barrier" in front of the first commercial sale of a
> Java product.  Sun got confused about its main goal: preventing
> Microsoft from putting Sun out of business by making the OS unimportant
> again, and got greedy just when a chance of success was looming large.]
>
>
> --
> Posted via Mailgate.ORG Server - http://www.Mailgate.ORG