comp.lang.idl-pvwave archive
Messages from Usenet group comp.lang.idl-pvwave, compiled by Paulo Penteado

Home » Public Forums » archive » Interesting Rant
Show: Today's Messages :: Show Polls :: Message Navigator
E-mail to friend 
Return to the default flat view Create a new topic Submit Reply
Re: Interesting Rant [message #51439 is a reply to message #51356] Fri, 17 November 2006 11:44 Go to previous messageGo to previous message
JD Smith is currently offline  JD Smith
Messages: 850
Registered: December 1999
Senior Member
On Thu, 16 Nov 2006 09:35:42 -0800, greg michael wrote:

>
>> [quoted text muted]
>
> I'm sure this is still backwards - this is how I see it:
> - vector operations are needed for their power of expression
> - IDL exists for this reason, and incidentally, makes them fast
> - loops turn out to be slow because IDL is interpreted
> - this is of secondary importance, because they're rarely necessary in

I'd love to see an overhead budget for a single trip around the IDL
interpreter loop. There are algorithms which no amount of cleverness can
recast into vector operations. For these, you can either code as a DLM,
or eat the horrible loop overhead. I've long argued for a "side loop"
capability of the language that would greatly reduce the
per-iteration overhead, at the cost of skipping processing of keyboard
input, widget events, etc., etc. Another (likely better) option would be
a more coherent interface to C code, i.e. make DLM writing more akin to
assembly writing, having it auto-compile, etc.

JD
[Message index]
 
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Read Message
Previous Topic: Re: Bitmap vectorization
Next Topic: Color background with PLOT

-=] Back to Top [=-
[ Syndicate this forum (XML) ] [ RSS ] [ PDF ]

Current Time: Sun Oct 12 06:44:16 PDT 2025

Total time taken to generate the page: 1.75254 seconds