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

rebol-philosophie - was "Any example of xml parsing on rebol site or reb

 [1/1] from: agem::crosswinds::net at: 6-Sep-2000 0:23


--- [youpi--technologies--wanadoo--fr] wrote on 5-Sep-2000/16:57:30+2:00
> Hello, > > - Is there any example of xml parsing somewhere since there seem to be > none on rebol.com or rebol.org ?
1) parse-xml or that?
> - Rebol doesn't seem to put accent on advanced concept of Object > Oriented Programming > like inheritance, polymorphism etc... is there any reason ? >
2) Oops? it has. is called 'object! , as everywhere.. but carl means, Rebol is advancter :) may be true, in Oops one has to declare _lots_ of pieces which can work together, but the "how" .. this is called patterns and _very_ modern. You can't express them in programming languages.. yes, forth'ees talking about 10 years longer, but they don't call it patterns and have'nt explaint it so systematic - maybe this systematic is an illusion.. but managers sleep better. there is another idea - 1) use fewer objects, but look heavy for similarities, like strings & blocks & - well, series. or protocolls :) and if you can recycle a relation (open it, put some stuff, close it. change position too, some searching) or (implement add and sub and mult and .. for ints, decimals, complex, matrices), do it. needs mutch of experience, i think.. 2) look at the human language. usually a good relation is here "encoded", if you use a similar structure for program as for thinking, it often works. (this is a Oops-hint). Carl says, a better way as Oops are dialects, or vocabularies, as forth calls it. in pattern-Oops you are allways implementing "lego-interfaces" to be able to plug pieces together. how to plug are informal rules in the comments ("never call this before open it!"..). a lot stuff to do: and hard to remember, it adds up.. but, more difficult, what are data and what are methods? when designing "i want to be able to tell this. i call it 'setHumba. hm. should the school get its children, or shall the children tell the school they are there. lets look in the pattern-book". a good forth/rebol-line can look like this: " peters child goto school ", in Oops you have to decide: "peter.getChild().goto(school)" or "school.requestKid(peter)" or .. this is equivalent for the user, but needs deep understanding by the programmer. And by the user, if he has to programm. in forth/rebol you write the plugging stuff in the grammar. in forth, the line above is more ideally, a bit data/action - order is usually requested. it can do deeper meta-stuff, but usually there are tricky side-effects if it gets a bit complicated. in rebol even this order should not be necessary, if i see carl's earlier fictive dialect samples. and with the block-parser, this may become true :) hm. now i browse to this interview..
> Thanks. >
-Volker