Re: Structures?
- Posted by Lucius Hilley III <lhilley at CDC.NET> Jan 19, 1999
- 520 views
On Tue, 19 Jan 1999 04:06:17 PST, stab master <stabmaster_ at HOTMAIL.COM> wrote: >Don't you think that this solution would be a little slow to do in >real-time? Yes. I wrote these incomplete routines for the ease of use of structures where structures maybe used quite frequently such as Windows programming. I wrote them simply to make the programming look more elegant. I am aware that they do suffer major speed problems. However, I have had intentions that IF I complete these routines the following features will also be included. A great Speed increase for both assign() and value_of(). Type checking of values that are stored within the structure. IE: It will be possible to create a structure that contains an atom or user defined number as the first element of a structure/sequence and a sequence as the element of the second structure. Granted this part is already possible but one would be capable of ensuring that a sequence is never written to the first element. >I mean, if one had a 3D-object defined by a structure: > >struct d3point { >int X >int Y >int Z >} > >(hope you understand C...) > > >Then you would have to call the assign and value_of routines in >real-time, wich I think would take way to much time to get any nice and >smooth graphics. I see you look to possibly make use of it in a high graphics medium, possibly a game. I will and do currently warn that these routines are dreadfully slow and should not be used in areas where speed is crucial/critical. I do plan to make the routines faster than they currently sit, If i finish it, But they aren't built-in as they should be and therefore can never be near fast enough for what is required for programming that relies heavily on high-speed. Built-in routines are definately the end result hoped for here. If enough people want these routines to exist then they will be completed. Also they, once completed, will be sent to Robert so that he can keep track of just how much these routines are wanted. If they are truely desired amongst many programmers then I feel it is possible for them to become built-in routines that would be fast enough for use with high-speed graphics. >I must admit that it could be easier (or it would look better anyway) to >use your struct routines instead of something like this: > >sequence point >point=repeat(0,3) >constant X=1 >constant Y=2 >constant Z=3 > >and then refer to the three space-coordinates as point[X], point[Y] and >point[Z]. But this "standard" solution would certainly be faster. > >But then ofcourse, maybe your routines weren't ment to be used for >things like that...(?) > > PS: Sorry about the novel above. Lucius L. Hilley III Dead Western Digital drive was Manufactured in August 98 and no longer functions. *UPDATE* is being replaced under warranty by Western Digital.