1  |   Order by index

Apr 23, 2003, 5:20pm

I don't know how this could be incorporated, but I was thinking about Cell
Space usage...

What if we could use shortcuts in the command line instead of completely
typing out the command?

create color = cc
create texture = ct
create sound = csound
create sign = csign
create solid = csolid
create name = cname
create noise = cnoise

Or, even finding a character that could be used in place of "create" ?

I know it doesn't seem like it could save that much space, but you'd be
amazed how many times
I have created a "control rod" (using create name) to be able to use the
same texture on various
objects in one area and it actually helped me save space!

Just a thought to possible get some ideas going?

Bastillion

Apr 23, 2003, 5:25pm

that would save tons of space. I have no idea why they haven't done it yet.

-Dion

[View Quote] What if we could use shortcuts in the command line instead of completely
typing out the command?

create color = cc
create texture = ct
create sound = csound
create sign = csign
create solid = csolid
create name = cname
create noise = cnoise

Or, even finding a character that could be used in place of "create" ?

I know it doesn't seem like it could save that much space, but you'd be
amazed how many times
I have created a "control rod" (using create name) to be able to use the
same texture on various
objects in one area and it actually helped me save space!

Just a thought to possible get some ideas going?

Bastillion

Apr 23, 2003, 5:39pm

Think about it... after the first create it actually takes more space.

- mark

[View Quote]

Apr 23, 2003, 7:05pm

If you were to use another create command, it wouldn't have to use the 'c'
before the name. So it would be like this:
"cc=red,texture=stone1"

-Dion

[View Quote] - mark

[View Quote]

Apr 23, 2003, 7:07pm

Wow, that's exactly the lines I was thinking!

Again, I don't have a clue how it would be implemented.. but that's why I'm
not a programmer. ;)

Bastillion

[View Quote]

Apr 23, 2003, 7:57pm

Exactly, but would need to maintain current 1s for it to work :)

- Mark

[View Quote]

Apr 23, 2003, 8:50pm

"strike rapier" <strike at rapiercom.freeserve.co.uk> wrote this
(news:3ea70c2f at server1.Activeworlds.com) in wishlist on 23 Apr 2003:

> Exactly, but would need to maintain current 1s for it to work :)
>
> - Mark
>
[View Quote] same with the colo[u]r commands and stuff

you would end up with something more along the lines of "c c red,t
av_skin9"

> completely
> be
> the
>
>
>

Apr 24, 2003, 4:01am

There was a similar idea that I like better just a bit above,
posted by Tomilius

Subject: Speed up the browser/world server

news://news.activeworlds.com/3e9c8b5c%241%40server1.Activeworlds.com


[View Quote]

Apr 29, 2003, 12:58am

Yep. Tokenizing the commands is an old idea, and a good one in my opinion. As for shortening them in text form, I suggested a long time ago that a system of bbreviation could be supported which would use the period character as a delimiter, and also that either the "create" trigger could be assumed when no trigger is given or the format of following a command name by a colon such as used to work with certain commands in older versions of AW could be supported on all commands.

Examples...

This example uses the implied create trigger...
activate sign "clicked"; sign "click here."

This example uses period delimited abbreviations...
c. co. red

This example uses the appended colon command syntax...
sound: forest1

For more information, see my November 23, 2002 post entitled "Ancient Wishes" news:3ddfffdd$1 at server1.Activeworlds.com

Here are some examples given in my Ancient Wishes" post...
c. Mu. 0 seed plant tree bigtree time=10
a. Mu. 1 chair time=60 name=object1, Mu. 100 onlight offlight time=1 name=object2
b. Tel. 30N 40W +1.2A +180
a. Tel. 3139N 106E +0; b. W. +0 +0 2.5A +180
c. Sc. 2 1 .5 time=50
b. Ac. +2S -1E

TechnoZeus


[View Quote]

Apr 29, 2003, 4:34am

Not exactly the same.

The tokenizing would take effect when the properties
are stored and expanded when you right-click a model.

It's the same method that VIC20, ZX80, Clipper and
older VB "compilers" used to speed up intepretation.
It replaces each keyword with a unique 1-byte token
that can be interpreted very quick at runtime.

Of course it would limit the keyword set to 255 commands
and 255 triggers - but with the current growth rate of
commands/triggers it will be sufficient for some time :)

It would even be helpful if one day someone would decide
to add a newbie mode with QBE for the action field.


Your wish is more an expert mode for the action field
representation that speeds up input of triggers/commands.

Both ideas would go very well together as just the array
with the text representation for each token would have to
be switched. Even "foreign" language command sets would
work. The tokens stay the same, the text representation
is customizable.


[View Quote]

Apr 30, 2003, 1:01am

It would appear you misread my statement. I said "...Tokenizing the commands is an old idea, and a good one in my opinion. As for shortening them in text form..." which it would seem either I didn't state clearly enough or you didn't read carefully enough... or both. :) I think it's a side effect of me writing so thoroughly that people end up skimming my posts. Hehe.

Actually, there's no reason why a two byte token system couldn't be used. That would allow the first byte to categorize the second byte, so that it would even be possible to distinguish between older and newer versions of the same command if a command's function changes drastically, such as what happened to the animate command back when Active World 3.0 came out. For now, it would make sense to simply have the first byte indicate whether the second byte should be expanded into a trigger, a command, or a parameter keyword such as "mask" or perhaps a value keyword such as "off" which would allow up to 256 possible in each category plus the option to add more than that through the simple addition of a category such as "additional commands" if necessary at some point in the future.


[View Quote]

Apr 30, 2003, 4:34am

Tokens => internal representation
Shortcuts => what you see in the build box

It's just a matter of separating function and design.


Btw.: Your posts are hard to read, as you do not
structure your text but :-/


[View Quote]

May 5, 2003, 3:15am

I apologize for my posts being hard for you to read. I try to make them as clear as I can manage, but different people have different ideas of what's easier to understand.

TechnoZeus

[View Quote]

Jul 7, 2003, 3:03am

A GZIP compressed stream essentially gives you multi byte tokenization
on *everything* free.

Anything AW can do to make the amount of data needed from the world
server more efficient is OK by me, but I don't really wanna know about it.

For example, the documentation insists that I should supply the .rwx at the
end of model names, so I do it even though it's easy to determine
empirically
that it is not really needed.

It may be however that it's not needed because the browser strips it off
when talking to the world server in order to save bandwidth.

The point being that whatever they do to compress the data stream should
be transparent to me.

[View Quote]

Jul 7, 2003, 10:43pm

It's not just about the data stream though. There are other reasons to tokanize commands.

TechnoZeus

[View Quote]

1  |   Order by index