Path: news.cs.au.dk!news.net.uni-c.dk!logbridge.uoregon.edu!skynet.be!poster!not-for-mail From: Atle Newsgroups: comp.lang.beta Subject: 'tokenize.bet' Date: Fri, 23 Jun 2000 12:21:00 -0100 Organization: Belgacom Skynet SA/NV Lines: 25 Message-ID: <3953643C.E1B51093@skynet.be> NNTP-Posting-Host: dialup280.charleroi.skynet.be Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: news1.skynet.be 961755367 30884 194.78.236.152 (23 Jun 2000 10:16:07 GMT) X-Complaints-To: abuse@skynet.be NNTP-Posting-Date: 23 Jun 2000 10:16:07 GMT X-Mailer: Mozilla 4.51 [en] (X11; I; Linux 2.2.5-15 i686) X-Accept-Language: en Xref: news.cs.au.dk comp.lang.beta:12481 Alllright: Next step. I have gotten a lot of help now, and I am wondering about how I would build a tokenizer. The exceptions that I have done like this, I know know how I can do nicer, but: I am heaviliy influenced by the first book I ever read on AI, which drew up 'robot arms' and told me that every part of the robot (finger, hand, etc.) should deal with normal execution and errors withing its own domain (NB! this is what I understood NB). IOW., it would be agains this philosophy to tell the 'arm' part about whether or not the object is 'held stable' - this is something the 'finger' would deal with. So, when I ask for a file in RW mode, I will let the 'open' part create the file if it does not exist. -- in this application, of course. I am not trying to establish any new kind of programming scheme here :-) But what I want to know now, is about the overall structure and use of the object FileScan, and some suggestions for building up an array (list?) of symbols, then a display of scope-visible symbols (2nd pass?) and suggestions for how to approach the problem of identifying what patterns an 'object assembler' may consist of. Thanks! Atle (PS I have a feeling my 'tokenizer' will be so ugly that many of you will close your eyes So please remember, I am learning every day :-) PS)