APL Killer App? 
Author Message
 APL Killer App?

Doens't APL seem like a natural for bioinformatics? One needs to access and
cross-reference huge amounts of data, the ability to do sophisticated
simulations, and crunch big numbers. It's also a very fast-moving and
dynamic target, so you have to be able to develop applications and modify
existing ones rapidly.

Anyone out there with a pile of VC? :)

--
-drl



Sun, 12 Sep 2004 07:13:13 GMT  
 APL Killer App?

Quote:

> Doens't APL seem like a natural for bioinformatics? One needs to access and
> cross-reference huge amounts of data, the ability to do sophisticated
> simulations, and crunch big numbers. It's also a very fast-moving and
> dynamic target, so you have to be able to develop applications and modify
> existing ones rapidly.

> Anyone out there with a pile of VC? :)

No VC, but this should be a good area for APL.  Python is currently
carving out a niche in this area.

Steve
--
"When I get a little money, I buy books; and if any is left
I buy food and clothes." - Erasmus



Sun, 12 Sep 2004 10:56:43 GMT  
 APL Killer App?

Quote:

>> Doens't APL seem like a natural for bioinformatics? One needs to access
>> and cross-reference huge amounts of data, the ability to do sophisticated
>> simulations, and crunch big numbers. It's also a very fast-moving and
>> dynamic target, so you have to be able to develop applications and modify
>> existing ones rapidly.

>> Anyone out there with a pile of VC? :)

> No VC, but this should be a good area for APL.  Python is currently
> carving out a niche in this area.

REALLY? I like Python, but frankly it is a toy. Then again, I hated Perl
with a mortal passion (oh my God what a hideous beast) and that became the
standard of CGI scripting, go figure.

Python is too abstract - it doesn't have any one thing it's targeted for.
APL on the other hand is a laser beam for numerics and interactive work.

--
-drl



Sun, 12 Sep 2004 20:02:52 GMT  
 APL Killer App?

Quote:

>> No VC, but this should be a good area for APL.  Python is currently
>> carving out a niche in this area.

> REALLY? I like Python, but frankly it is a toy. Then again, I hated Perl
> with a mortal passion (oh my God what a hideous beast) and that became the
> standard of CGI scripting, go figure.

Funny you should mention Perl

- it's become the tool of choice for bioinformatics, I've never even heard
of someone using python in that field....

See "Beginning Perl for Bioinformatics" by James Tisdall.

http://www.perl.com/pub/a/2002/01/02/bioinf.html
http://www.oreilly.com/catalog/begperlbio/

I think one of the biggest problems Perl has is getting people to stop
thinking about it as a CGI scripting language -  there's lots of
*really* *bad* "cgi-perl-for-dummies" style books that DON'T teach you Perl,
but teach you crappy BASIC-style code  with a keyword search/replace to
make it look like Perl...  The code will run, but it's far from idiomatic
Perl -  Kinda like painstakingly coding lots of explicit loops over arrays
in APL...

The only sensible way to learn proper Perl is from the O'Reilly books.
(Perl actually being a {*filter*} by O'Reilly and Larry Wall to sell the
books, of course... :-) )

And then there's PDL - it adds some multimentional matrix goodness
to Perl.  It's a toy by modern APL-family standards, but compares reasonably
well to Matlab.  

http://pdl.perl.org/

'PDL ("Perl Data Language'') gives standard Perl the ability to compactly
store and speedily manipulate the large N-dimensional data arrays which are
the bread and butter of scientific computing.'

--
Don't eat yellow snow.



Mon, 13 Sep 2004 05:04:10 GMT  
 APL Killer App?

Quote:

> <snip/>
>>No VC, but this should be a good area for APL.  Python is currently
>>carving out a niche in this area.

> REALLY? I like Python, but frankly it is a toy. Then again, I hated Perl
> with a mortal passion (oh my God what a hideous beast) and that became the
> standard of CGI scripting, go figure.

> Python is too abstract - it doesn't have any one thing it's targeted for.
> APL on the other hand is a laser beam for numerics and interactive work.

Python is hardly a toy, though it lacks APL's focus.  It is being used
for serious "numeric"://http://www.pfdubois.com/numpy/ work with the
appropriate extension modules.

Steve
--
"When I get a little money, I buy books; and if any is left
I buy food and clothes." - Erasmus



Mon, 13 Sep 2004 10:14:01 GMT  
 APL Killer App?

Quote:

> <snip/>

> Funny you should mention Perl

> - it's become the tool of choice for bioinformatics, I've never even heard
> of someone using Python in that field....

Perhaps not "tool of choice" yet, but the "Biopython
Project":http://www.biopython.org" has some resources.  A Google search
shows many more.

Steve
--
"When I get a little money, I buy books; and if any is left
I buy food and clothes." - Erasmus



Mon, 13 Sep 2004 10:27:01 GMT  
 APL Killer App?
                              Re: APL Killer App?
Quote:


>   Reply to: Danny Ross Lunsford
>   Date: Wed, 27 Mar 2002 12:02:52 GMT
>   Organization: 1445
>   Newsgroups:
>          comp.lang.apl


>>> Doens't APL seem like a natural for bioinformatics? One needs to access
>>> and cross-reference huge amounts of data, the ability to do sophisticated
>>> simulations, and crunch big numbers. It's also a very fast-moving and
>>> dynamic target, so you have to be able to develop applications and modify
>>> existing ones rapidly.

>>> Anyone out there with a pile of VC? :)

>> No VC, but this should be a good area for APL.  Python is currently
>> carving out a niche in this area.

>REALLY? I like Python, but frankly it is a toy. Then again, I hated Perl
>with a mortal passion (oh my God what a hideous beast) and that became the
>standard of CGI scripting, go figure.

>Python is too abstract - it doesn't have any one thing it's targeted for.
>APL on the other hand is a laser beam for numerics and interactive work.

    I have written a few functions to scroll ascii drawings of dna and
    rna molecules. This can be downloaded from
    http://d4maths.lowtech.org/mirage/install.htm

    If you run windows systems download DNA.EXE.
    For linux get math2545.tgz, and compile it.

    load the application ya, by typing the following at the dollar prompt.

    ld ya.d4f
    ya_ss

    Tony Goddard

    The same files also contain scripts for browsing images on the computer,
    and much else besides.



Tue, 14 Sep 2004 01:48:58 GMT  
 APL Killer App?
Nice! Are you heavily into symbolics? I want to start on a computer algebra
system for APL and would welcome suggestions and pointers to existing work
that is considered good.

-drl

Quote:

>     I have written a few functions to scroll ascii drawings of dna and
>     rna molecules. This can be downloaded from
>     http://d4maths.lowtech.org/mirage/install.htm

>     If you run windows systems download DNA.EXE.
>     For linux get math2545.tgz, and compile it.

>     load the application ya, by typing the following at the dollar prompt.

>     ld ya.d4f
>     ya_ss

>     Tony Goddard

>     The same files also contain scripts for browsing images on the
computer,
>     and much else besides.



Tue, 14 Sep 2004 09:48:39 GMT  
 APL Killer App?
Ugh!

Have you EVER tried to maintain someone else's Perl? It truly can become
"write-only" code. I never saw even bad APL that was as hard to read as bad,
uncommented Perl - and the prefix syntax does not help. I simply can't
imagine Perl for something as complex as BI.

Perl - a horse designed by a committee of horses. :)

-drl


Quote:

> >> No VC, but this should be a good area for APL.  Python is currently
> >> carving out a niche in this area.

> > REALLY? I like Python, but frankly it is a toy. Then again, I hated Perl
> > with a mortal passion (oh my God what a hideous beast) and that became
the
> > standard of CGI scripting, go figure.

> Funny you should mention Perl

> - it's become the tool of choice for bioinformatics, I've never even heard
> of someone using Python in that field....

> See "Beginning Perl for Bioinformatics" by James Tisdall.

> http://www.perl.com/pub/a/2002/01/02/bioinf.html
> http://www.oreilly.com/catalog/begperlbio/

> I think one of the biggest problems Perl has is getting people to stop
> thinking about it as a CGI scripting language -  there's lots of
> *really* *bad* "cgi-perl-for-dummies" style books that DON'T teach you
Perl,
> but teach you crappy BASIC-style code  with a keyword search/replace to
> make it look like Perl...  The code will run, but it's far from idiomatic
> Perl -  Kinda like painstakingly coding lots of explicit loops over arrays
> in APL...

> The only sensible way to learn proper Perl is from the O'Reilly books.
> (Perl actually being a {*filter*} by O'Reilly and Larry Wall to sell the
> books, of course... :-) )

> And then there's PDL - it adds some multimentional matrix goodness
> to Perl.  It's a toy by modern APL-family standards, but compares
reasonably
> well to Matlab.

> http://pdl.perl.org/

> 'PDL ("Perl Data Language'') gives standard Perl the ability to compactly
> store and speedily manipulate the large N-dimensional data arrays which
are
> the bread and butter of scientific computing.'

> --
> Don't eat yellow snow.



Tue, 14 Sep 2004 09:52:31 GMT  
 APL Killer App?

Quote:

> Ugh!

> Have you EVER tried to maintain someone else's Perl? It truly can become
> "write-only" code.

It can indeed - particularly if the someone else is writing
perl-as-if-it-were-some_other_language - having a hairy
loop where "map" or "grep" would do fine, and so on.  Or, of course, if
someone is misguidedly "showing off" by writing obfuscated code.

OTOH, exactly the same argument applies to Perl as to APL - once
you learn to read the language, which might take a little longer, than,
say, VB, it starts to make much more sense.   One of Perl's "problems"
is definitely shared by some ASCIIfied APL-derived languages - the ASCII
symbols just aren't as nice as APL's custom symbols and are already far too
overloaded with other meainings from other contexts, and thus make the
source code look like line noise.

One "solution" that some Perl people use, thanks to Perl's mutable syntax,
is "English" modules and so on*, that seem to have the net effect of making
Perl look more like COBOL (Arrgh!) - but for true, long-term,
learned-the-language readability, it might be better to go the other way,
brteaking away from ascii and use Unicode identifiers... which one can
AFAIK do in Perl - one could even conceive of defining an anaemic (no
nesting), but syntax-and-symbol-compatible APL on top of PDL..  

The problem there is that input methods have by and large not caught up
with character set encoding technology -  Even implementations of the
obvious one, handwriting recognition, still seem flawed because
implementors seem to stop (at least in the western world), once they've
got translation working for all the characters that are already on the
standard PC keyboards anyway -  there's no good technical reason
why handwriting recognition programs couldn't spit out the Unicode-encoded
APL symbols when you wrote them...

Or maybe there are keyboards with  keytops that are programmable LCD
displays.  That'd be nice too.   :-).

* have a look at Lingua:Perligata:Romana some time for a related idea -
makes perl read like Latin.
http://www.csse.monash.edu.au/~damian/papers/HTML/Perligata.html

--
Don't eat yellow snow.



Tue, 14 Sep 2004 22:20:58 GMT  
 APL Killer App?

Quote:
> Doesn't APL seem like a natural for bioinformatics?

From a computing point of view, the bioinformatics community is very
conservative IMO. They go with the {*filter*} stuff. Furthermore Perl is
already entrenched (and yes I know about the BioPython and BioJava
projects but if you look at the job ads, for data-crunching stuff
the skills they want are C/C++ and perl) and non-commercial code is
shared a lot, which automatically marginalizes you and your code if
you write in an {*filter*} language.

That's not to say that a very productive and politically astute fellow
could not succedd in carving a niche for APL (after all, perl had to
start from the ground floor, too) but that it is unlikely to happen.
IMO if you wanted to carve a niche for APL in biology, you'd have to
start at a lower echelon, i.e., teach APL to biology undergrads and
hope some of them keep using it once out of school. Given how computer-
phobic many biologists (and doctors!) are, APL might appeal to them.

-- O.L.



Wed, 22 Sep 2004 03:09:09 GMT  
 APL Killer App?

Quote:

> > Doesn't APL seem like a natural for bioinformatics?

> From a computing point of view, the bioinformatics community is very
> conservative IMO. They go with the {*filter*} stuff.

This doesn't make sense. I was expecting that your second sentece would be
"they still use fortran 66" or something similar.

Bye, Dragan

--
Dragan Cvetkovic,

To be or not to be is true. G. Boole



Wed, 22 Sep 2004 03:12:13 GMT  
 APL Killer App?

Quote:
>> From a computing point of view, the bioinformatics community is very
>> conservative IMO. They go with the {*filter*} stuff.

> This doesn't make sense. I was expecting that your second sentece would
> be "they still use FORTRAN 66" or something similar.

Going with the latest popular language can be a conservative attitude:
there's saftey in numbers, or so the reasoning goes. For one thing, it
looks better on your CV, should you ever want to switch fields.

-- O.L.



Fri, 24 Sep 2004 22:44:30 GMT  
 APL Killer App?

Quote:

> Going with the latest popular language can be a conservative attitude:
> there's saftey in numbers, or so the reasoning goes. For one thing, it
> looks better on your CV, should you ever want to switch fields.

Yes! Example, C++ - a horrible implementation of OOP (although a useful
upgrade to C if used correctly). Smalltalk would have been my OOP of choice,
but that would just be too radical for the mainstream.

Come to think of it, Smalltalk and APL lack general popularity for the very
same reason.

-drl



Sat, 25 Sep 2004 18:57:27 GMT  
 APL Killer App?
OL>  
OL>  IMO if you wanted to carve a niche for APL in biology, you'd have to
OL>  start at a lower echelon, i.e., teach APL to biology undergrads and
OL>  hope some of them keep using it once out of school. Given how computer-
OL>  phobic many biologists (and doctors!) are, APL might appeal to them.

We did that in the mid-80ies:  introductory programming courses in
APL for non-CS (and CS, too) students, some of them majoring in
biology and otherwise not related to computer subjects.  The course
was well received BUT:

APL didn't spread around much further in the bio depts, nor did we
have more bio students coming to us in droves.  I would explain
this with 2 reasons:

(1) Since those students were *real* programming beginners they were
    unable to compare APL's ease to the pain which all the non-APLers
    have to bear.  So there was no point for them to recommend
    particularly APL to their fellow bio students.

(2) Yes.  That special keyboard thing made it impossible to continue
    with APL outside of our course, for example in a bio lab with
    standard equipment.

                                                        Martin Neitzel



Sat, 25 Sep 2004 07:44:41 GMT  
 
 [ 20 post ]  Go to page: [1] [2]

 Relevant Pages 

1. Updating APL and Killer Apps

2. Web Apps as ST's Killer

3. Killer apps -- for n'th time

4. A Diffident Killer App

5. A Forth Killer App

6. Forth Killer App

7. FORTH's killer app?

8. killer app the n-th ?

9. Killer apps - revisited

10. Forth "killer app"

11. Forth "killer app"

12. Killer app for Ruby developers?

 

 
Powered by phpBB® Forum Software