I just saw this thread and I thought I would put my two cents in. I use Unicon for searching large texts and reformatting them occasionally.
When it comes to text manipulation no other language is better than the Unicon/Icon/Snobol family. But I think Ruby is popular because its syntax is elegant, because it is entirely object oriented, like small talk, because it is a scripting language like that, and because it has some text processing features built in like some of the other scripting languages. I would like to see some marriage between Ruby and Unicon. The idea of making Unicon a DSL on top of Ruby would be very appealing to me. But I don't know what that would do for performance. Both Icon and Ruby have C interfaces. Perhaps there is some way to marry them via their interfaces. Is that possible? I talked to Clint about XML and Unicon several years ago. It seems to me the advent of XML was a lost opportunity for Unicon. Unicon is obviously the best way to deal with XML, but somehow no one recognized this fact in the larger world. Now I think Ruby has displaced any possibility for Unicon to become popular because it has enough text processing power to make it a good candidate for that use, plus it has all the other features that makes it desirable for programmers who believe in object oriented programming. I think the object model in Unicon, just like the functional model in Icon, something added on to the original language idea in Snobol The problem is that there are interference between these ideas and the basic idea of Snobol. What was needed was a rethinking of the way that the functional and object models relate to the fundamental concept of snobol, and the production of an elegent language that promoted all these various programming ideas without their interfering with each other. When I write Unicon programs I don't use either the functional or the object ideas, but rather use the string searching directly, i.e. I use the Snobol Core of the language. It is usually very small programs I am writing to get at some aspect of the text I am processing that I cannot get to in any other way. This sting searching core, which goes beyond what is available in other languages is precisely what is needed for XML applications or anything that has to do with complex string searching within texts or parsing. If that was made available in RUBY then that would be excellent. I don't think the folks who are into Ruby know that there is more to string searching than what they have currently. I would really have liked to see Unicon in the position that Ruby is in now, but with the advent of Ruby there is no chance of that. Ruby has one of the most sophisticated and elegant syntaxes of any programming language to date. That is why people like it so much. The Uncon syntax is not that elegant, and so despite its power for parsing and text manipulation, it cannot have wide appeal. It's power is too specialized. And everything else it does is just like other languages. Ruby on the other hand sets itself apart from other languages by it over all sophistication and elegance of syntax in every aspect of the language. Anyway this is my two cents on this subject. Sorry it took me so long to see this thread and respond to it. Kent Palmer [EMAIL PROTECTED] http://think.net At 07:05 AM 10/6/2006, you wrote: >On 6 Oct 2006, at 15:51, Steve Wampler wrote: > > I agree! This would be nice. Every so often I think that I'd like to > > get back into working on some implementation issues, but the current > > code hurts my brain. I understand the motivations behind the > > current solution, but that doesn't keep my brain from hurting. > > (I also agree that such a redesign would be, uh, non-trivial.) > >Yes, I tried getting into the JCon code during one of my brief >infatuations with Java, but that was just as bad. >Another idea that might be fun (but not even vaguely justifiable) >would be to implement Icon as a DSL in Ruby, and then build Unicon on >top of that. I find the idea of running up an Unicon application >inside Rails sick enough to be appealing >8D > > > Some form of threading ('live co-expressions', perhaps?) would be > > wonderful. I'd like to see it designed right, however, and not just > > another "let's throw POSIX threads in" solution. It would be fun > > to talk about possible approaches and what would fit well in Unicon. > >That's the biggest problem - how to do it naturally. I don't really >like the threading models most languages use anyway, as they're too >concerned with flow of control and not enough with generation of >expressions - perhaps the best approach would be to extend the >monitoring/profiling extensions in some way although I've not played >with them enough to be certain. Or else some kind of parallel >generator which kicks out more than one result at the same time but >thinking about that's making my brain feel weird. > >Ellie > > >Eleanor McHugh >-- >http://eleanor.goth-chic.org/ >http://feyeleanor.livejournal.com/ > > > >------------------------------------------------------------------------- >Take Surveys. Earn Cash. Influence the Future of IT >Join SourceForge.net's Techsay panel and you'll get the chance to share your >opinions on IT & business topics through brief surveys -- and earn cash >http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV >_______________________________________________ >Unicon-group mailing list >Unicon-group@lists.sourceforge.net >https://lists.sourceforge.net/lists/listinfo/unicon-group ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys-and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ Unicon-group mailing list Unicon-group@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/unicon-group
