MOO-cows Mailing List Archive
[Prev][Next][Index][Thread]

Re: verbing properties again

On Fri, 26 Jul 1996, Kipp the Kidd wrote:
> At 10:44 AM 7/26/96 PDT, Adams, Charles wrote:
> >If I want to access raw data, I want the raw data, period.  None of this 
> >"well you really want to do this so I'll do it instead" junk.
> 
> One question -- WHY???  The raw data belongs to the object.  It might not
> even be accurate if a verb is supposed to be handling it.  Alot of times, as
> previously mentioned, _msgs are set to lists and the verb picks a random
> element.  The raw data will break the verb that tries to display it.  It is
> rare to really need raw data.  In fact, it is so rare, there is only one
> verb in the Paradigm core that actually needs raw data.
> 
> - Kipp
> http://radioactivo.com.mx/~paradigm

Why? Because it is convinient, and when the raw data is what you get from 
the verb that returns the value, 
@program #1:get_*
":get_<propname> => returns the value of this.<propname>";
"Or....some modified value from some object redefining this verb.";
set_task_perms(caller_perms());
return `val = this.(verb[5..$]) ! ANY => val';
.

Then, you might as well get the raw data.

And as mentioned numerous times, if you realy want some property access 
through a wrapper verb, make the property !c!r!w, and provide verbs for 
reading and setting it.

And talking about making the MOO programming language more "real" OO, 
then you can try to convince us all which OO theory is the best.. How 
about methods...tasks communicating through messages only..(send and 
receive) or direct/inderict function calls. Try to concvince my old 
professor, Ole Johan Dahl, who wrote Simula, that no direct access should 
be allowed. Try with the C++ makers.. 


Someone would say "real" OO is each verb on different objects, will only 
be allowed to communicat through a send and receive of messages, which 
means, each object would be a running task, waiting to receive a message. 
This is, form what I know, some of the new theory on concurrent OO 
programming. And of course, everything is an object. Down to numbers, 
letters, and everything has methods to call, depending on the message 
they receive. To add two numbers, send a message to one of the numbers, 
ask it to add itself and this other thing, which would be the "address" 
for where to send a message to the other number and then send you the 
answer when done.


And in some previous post, someone asked how boring description would be 
if it was always the same, with just a raw property reading. But isn't 
that why we, at least in LambdaCore, has look calling look_self calling 
description reading .description? Lots of places to put in modifications 
here. Just that you can read the property doesn't mean you will looses 
something. Then you would just have to take away all "non real" features 
in MOO. Make it look exactly like "real life". Where all you could see of 
someone is what they show you. You can not look at how people realy look 
when you meet them on the street. At least not without getting arrested.


Well. Maybe the soloution is to make a new language. Not based on the 
exsisiting MOO language. With different forms of communications..and 
different behaviours. Then you could get what you want, and feel is a 
"real" OO programming language.

And I would like a language like that actually. Makes structured 
programming easier to learn for others. But, I do not think LambdaMOO is 
the right thing to turn into something like that. Maybe LambdaMOO++ :)

And I hope you have all seen what happens when you take something that 
has been around for a long time, C, and turns it into something new, C++. 
That is a bad idea, if you ask me. (Not that anyone will..)

Oh well, enough for now,
  Kristian.
References:
Home | Subject Index | Thread Index