mo@messy.bellcore.com (Michael O'Dell) (05/29/90)
If the concerns about Postscript is that is is hard to write en masse, then having a C compiler which generates Postscript should alleviate a lot of those concerns (the meta-discussion of whether C is a programming language as well isn't appropriate for this group, and is moot as well). Further, if one looks closely at the existing NFS protocol and the NeFS built-ins, one will discover that it takes very little Postscript to implement most of what you really want to do. By all means, a certain amount of reason should prevail, like: I would not recommened an attempt to implement a database system in Postscript so it can be downloaded to the server!!! But by the same token, I have read (but admittedly not written much) NeWS code done with the Class system, and it certainly doesn't read as awful as some would imply. Have people been badly burned and are speaking from hard-won experience here? -Mike
pallas@Neon.Stanford.EDU (Joe Pallas) (05/30/90)
In <23683@bellcore.bellcore.com> mo@messy.bellcore.com (Michael O'Dell) writes: >But by the same token, I have read (but admittedly not >written much) NeWS code done with the Class system, and >it certainly doesn't read as awful as some would imply. The NeWS Class system is best described as an example of an object-oriented language implemented in PostScript. Granted, it is an embedded implementation, so all of PostScript is exposed to the programmer. But its very existence is evidence of PostScript's inadequacy as a programming language. >Have people been badly burned and are speaking from hard-won >experience here? I know I have and am. I don't know about anyone else. joe
mh@awds26.imsd.contel.com (Mike Hoegeman) (05/31/90)
In article <1990May30.030703.462@Neon.Stanford.EDU> pallas@Neon.Stanford.EDU (Joe Pallas) writes: > >The NeWS Class system is best described as an example of an >object-oriented language implemented in PostScript. Granted, it is an >embedded implementation, so all of PostScript is exposed to the >programmer. But its very existence is evidence of PostScript's >inadequacy as a programming language. > Well... you can make this rather flippant comment about almost any programming language. For example: "The very existence of the C standard library is evidence of C's inadequacy as a programming language." Which is a pretty ridiculous statement. I've used PostScript quite a bit (probably more than most of the PostScript flamers in previous postings) and i find it similar to lisp in ease of use , maintainablilty, etc.. >>Have people been badly burned and are speaking from hard-won >>experience here? > >I know I have and am. I don't know about anyone else. How have you been burned? This is, for the most part, genuine curiousity. I'd be interested in hearing something more specific. I don't have an opinion on NeFS one way or the other yet but this "I don't like PostScript, it's icky" commentary is non-informative at best.
aperez@cvbnet.UUCP (Arturo Perez x6739) (05/31/90)
From article <53478@wlbr.IMSD.CONTEL.COM>, by mh@awds26.imsd.contel.com (Mike Hoegeman): > In article <1990May30.030703.462@Neon.Stanford.EDU> pallas@Neon.Stanford.EDU (Joe Pallas) writes: > > > >The NeWS Class system is best described as an example of an > >object-oriented language implemented in PostScript. Granted, it is an > >embedded implementation, so all of PostScript is exposed to the > >programmer. But its very existence is evidence of PostScript's > >inadequacy as a programming language. > > > > Well... you can make this rather flippant comment about almost any programming > language. For example: > > "The very existence of the C standard library is evidence of C's > inadequacy as a programming language." > > Which is a pretty ridiculous statement. I've used PostScript quite a bit > (probably more than most of the PostScript flamers in previous postings) > and i find it similar to lisp in ease of use , maintainablilty, etc.. > I don't think this scans in quite the way that Joe Pallas' statement did. A more appropriate statement would have been: "The very existence of the X toolkit and widgets is evidence of C's inadequacy as a graphical (or UI) programming language." Anyway, all I know about Postscript is that it's ugly and proprietary (not necessarily the same thing :-). What makes Postscript so hard (or easy) to program in? And why should we even want to write filesystems in a language designed for page description? Why isn't someone proposing a language designed for something a little more general? For example, how about Smalltalk? Hey, just 'cause Postscript is out there doesn't mean we have to use it. We could probably use any interpretable language, Lisp, Forth, or even Perl! Again, why should we use a page description language for a filesystem? Arturo Perez ComputerVision, a division of Prime aperez@cvbnet.prime.com Too much information, like a bullet through my brain -- The Police
pallas@Neon.Stanford.EDU (Joe Pallas) (06/01/90)
In <53478@wlbr.IMSD.CONTEL.COM> mh@awds26.imsd.contel.com (Mike Hoegeman) writes: [quote of my saying the NeWS Class system is evidence of PostScript's inadequacy] >Well... you can make this rather flippant comment about almost any programming >language. For example: > "The very existence of the C standard library is evidence of C's > inadequacy as a programming language." >Which is a pretty ridiculous statement. Yes, I agree---because using standard library does not change the fundamental model of computation the way the NeWS objects-in-PostScript system does. The fundamental operation in the NeWS class system is "send message", which is not a PostScript operation. The fundamental operation in the use of a C library is "call function", which is a C operation. But this is not worth arguing over. >How have you been burned? This is, for the most part, genuine >curiousity. I'd be interested in hearing something more specific. I've been burned by dynamic binding, which makes it practically impossible to write recursive procedures with named own variables. I've been burned by stack overflow/underflow because some procedure didn't leave the stack the way some other procedure expected. I've been burned by STATIC binding, because Adobe encouraged people to use the "bind" operator, which improves performance by completely changing the semantics of the language. I've also been burned by the machine-dependent limit on the length of an executable array literal, when I had some code being generated by a program. All of this is just the language, excluding the times I've been burned by machine-dependent graphics operations. joe
Dan@dna.lth.se (Dan Oscarsson) (06/03/90)
In article <506@cvbnetPrime.COM> aperez@cvbnet.UUCP (Arturo Perez x6739) writes: > >Anyway, all I know about Postscript is that it's ugly and proprietary (not >necessarily the same thing :-). > >What makes Postscript so hard (or easy) to program in? And why should we even >want to write filesystems in a language designed for page description? > >Why isn't someone proposing a language designed for something a little more >general? For example, how about Smalltalk? > >Hey, just 'cause Postscript is out there doesn't mean we have to use it. We >could probably use any interpretable language, Lisp, Forth, or even Perl! > PostScript is nice and easy to program in. Smalltalk, Lisp, Forth are ugly. I would never think of using Smalltalk or Lisp for a NeFS or NeWS server. Dan
alpope@skids.Sun.COM (Alan L Pope) (06/05/90)
In article <506@cvbnetPrime.COM>, aperez@cvbnet.UUCP (Arturo Perez x6739) writes: > > Hey, just 'cause Postscript is out there doesn't mean we have to use it. We > could probably use any interpretable language, Lisp, Forth, or even Perl! Yeah. Great. Let's use TECO. Alan Pope <alpope@Eng.Sun.COM>
tok@stiatl.UUCP (Terry Kane) (06/06/90)
Dan@dna.lth.se (Dan Oscarsson) writes: >PostScript is nice and easy to program in. Smalltalk, Lisp, Forth are ugly. And Dick Nixon was not a crook. :-) -- Terry Kane gatech!stiatl!tok Sales Technologies, Inc 3399 Peachtree Rd, NE Atlanta, GA (404) 841-4000