On Mon, 13 Mar 2000 07:16:50 -0500, Chris Huff
<chr### [at] yahoocom> wrote:

>The reason it got a consistently negative response is that it would 
>require a graphical editor to comprehend the simplest scene written in 
>that language. 

Perhaps, for you. I think that the vast majority of people would prove
your statement grossly exaggerated at best. Your statements carry the
tone of a person who is either afraid of a new idea, or doesn't
understand it, and vehemently scorns any mention of it. In fact, you
have repeatedly said that you didn't understand the benefits that
Nigel has listed for you,  but you seem to still feel qualified to
dismiss the idea as impractical. How can you properly judge that which
you don't understand?

I, on the other hand, would actually find it easier to read and write
in an XML style. Oh sure, it's more typing (assuming you don't have an
advanced editor), but it is more logical, predictable, and extensible.
The "punctuation" is also *MUCH* simpler with tags, because you don't
have to remember whether to use parenthesis, brackets, braces, single
quotes, or double quotes  with a given keyword. For the most part,
there are only <> characters for punctuation.

Someone has said that the braces in POV make it really clear and easy
to understand. I say that closing tags, which are more specific than
simple closing braces, are even more simple to sort out. When I see a
closing brace, I have to back track and count braces to see what
keyword that closing brace is associated with. If it were a closing
TAG, then I would know immediately which keyword it was associated
with. How many times have you wasted time chasing a missing closing
brace, or perhaps had one brace too many? (Everyone has an occaisonal
typo, everyone from Ken to Ron Parker.) It would be much quicker to
find a missing or extraneous tag, because they are directly labeled as
to *what* they are actually closing.

>there would be no reason to use it

We've already given you a few reasons to use it, so that comment is
totally incorrect. I can only assume you mean it euphemistically.

>, and several reasons not to use it

Even though I am intrigued by the idea of POVML (XML styled POV), I
will agree that it would be more typing, at least in its *raw form*. I
have suggested in the past that if POVML were adopted, it might be a
good idea to create a cross-platform editor for it that tokenizes the
raw form to conserve file space. The idea is similar to what Atari
BASIC did. You typed in the BASIC program, and the BASIC interpreter
tokenized your file into a much more compact form for its internal
use. The end user only saw the expanded, human readable form, but the
file was actually stored on disk in a tokenized form. This
tokenization is not to be confused with the language interpretation
that happened at run time. In some respects, Atari BASIC was somewhere
between a pure interpreted, and a pure compiled language. Doing
something like this for POVML would solve the filesize problem, and
probably shorten the render-time parsing a little to boot, since at
least some of the parsing work could be done at entry time, during the
tokenization.

This editor could also automatically supply closing tags for you when
you typed the original tag. That would eliminate half the added typing
right there. It would also insure that you always *had* a closing tag,
and never forgot to type one. (Unlike the current situation with
braces and our current editors, where one can easily lose count and
forget a closing brace in a complicated piece of code.)

I think that XML styled POV code deserves exploration.

Later,
Glen Berry

Post a reply to this message