Mailing List Archive: 49091 messages
  • Home
  • Script library
  • AltME Archive
  • Mailing list
  • Articles Index
  • Site search

[REBOL] Re: Project introduction ... Re: Re: RFC: Rebol Framework

From: al:bri:xtra at: 9-Oct-2002 16:48

pekr wrote:
> I meant something bit different - not real rebol objects, but e.g. earlier
mentined accommodation class. System admin (not necessarily rebol programmer) can say to himself one day - hey, this hotel has cinema inside. Where should I put that? So he goes to "object" repository, selects hotel class, and adds/describes "attribute" (field) of 'cinema-inside? type e.g. That was what I meant by ability to extend "object" during the system life-time ...Of course there can be consequences to it, like such field not existing in former records - so ti may cause some difficulcy to query engine e.g. I want and need users to be add or remove fields. A fixed system won't be better than what we've got all ready.
> > Because one can alter the block at runtime! For example, adding more
fields or removing them. That allows customising each record.
> You can do so with object too. Object = template you load from your
repository. If you want to extend, you can extend the class (block of set-word and value pairs), and instantiate the object (load block into context) once again :-) Unfortunately, one has to then hunt for the other instances of the object in the net and delete them. It's far more simple and quicker to just clear the block! (if necessary) then add in fields (Name Value pairs) as required. Sure, if you're just storing a tree of objects your system will work, but not if one has network of values. Andrew Martin ICQ: 26227169