|
 |
Christopher James Huff wrote:
> In article <3d895e0d@news.povray.org>,
> David McCabe <dav### [at] myrealbox com> wrote:
>
>> 1. The user gets a real scripting language that can do any kind of
>> calculations or functions or whatever, and can even integrate
>> rendering with other tasks, automating something or other.
>
> A "real scripting language"? How is POV not "real"?
> Also, there are security issues...if it can do anything, it can do
> ANYTHING.
POV is not real...because it can't do everythingor it makes it really hard.
POV is like PostScript, which you wouldn't exactly call a scripting
language, even though it can do that kind of thing to some degree.
Security is an issue whever you are writing a program. How is writing a
script for POV less secure then writing a script that runs alone?
>
>
>> 2. The user doesn't have to learn a new language, just a new API.
>
> You assume the user already knows the language. Most users of POV don't
> have any programming experience other than POV.
OK.
>
>
>> 3. You get a high-quality interpeter for free. This would greatly reduce
>> the size of POV-Ray, etc.
>
> How would it make POV smaller? Instead of POV interpreter + POV core,
> you end up with POV core + generic interpreter + lots of glue code.
No, the interpreter would be an *external dependency* (get used to it :-).
>> But I digress. Of course, that which I have just briefly described could
>> be accomplished using a specialized language. The advantage of a normal
>> language is that everyone gets stuff for free.
>
> The disadvantage: you get what you get. You can then mutate it into a
> custom language, making it more suited for your purposes but harder to
> update and losing compatibility with the "official" language, or you can
> stumble along with the unmodified version.
Sure. That is a disadvantage, and a really big one too, such that it really
kills the idea.
>> Anyway, my nominations for the language are (in this order):
>>
>> 1. Ruby
>
> Not really designed to be embedded in another application, possible to
> do but awkward. I think Matz is working on making this easier.
it's a good language, however.
>> 2. Python
>
> ***gags***
Here, here! I would prefer Java 1000 times, but the syntax would be *so*
clunky for an SDL application.
>> 3. Java
>
> Most likely for a programmer to know, probably the fastest, but I'd hate
> to use it for scene description.
So would I, or it would be nomination #1.
>
>
> As someone else mentioned, it has been discussed many, many times
> before,
Pordon me. I'm new.
> but I'll say this anyway: the advantage of POV-script is that it
> is built for scene description. For example: in those languages you
> either have to create a class for vectors (which makes using them clumsy
> and takes lots of typing for the simplest things) or give up vectors
> completely, using the terrible solution of specifying 3 floats all the
> time (making it hard to tell a vector from a bunch of floats and making
> vector math nearly impossible, making bugs in math code almost certain
> for things that are extremely simple in POV).
Yeah.
>
> Also, you get a pre-built language that is more powerful than the
> existing POV language, which is also harder to use for the simplest
> scene description tasks which don't require any programming and harder
> for non-programmers to learn. I started CSDL/Sapphire as a scene
> description language, and though I now think it is too complex for this
> use, I think it would be far better than any of the languages you
> mentioned simply because it supports vectors.
OK.
> I think any new scene description language will have to be something
> like POV: partly a simple markup language or human-editable data format,
> with a scripting language and built-in support for vectors. Object
> oriented as well...POV is almost there, it could become a
> prototype-based OOL with just a few modifications.
Okay, that's fine. But please make sure we get good OO features.
--
David McCabe
http://12.225.144.95/pgpkey
Post a reply to this message
|
|
