hadden@ella.SRC.Honeywell.COM (George D. Hadden) (04/18/89)
a piece of information: the prolog on simtel (version 1.90n) is
broken. here is one manifestation: if you define append as follows:
append([],L,L).
append([X|L1],L2,[X|L3]) :- append(L1,L2,L3).
and then type
?- append(X,Y,[a,b]).
you will get
X=[], Y=[a,b]
X=[a], Y=[b]
X=[a,b], Y=L_12
obviously, the last should be Y=[]. i've called a.d.a. about it and
they say that the newer version doesn't have this problem. i'm not
sure what the mechanism is to get this stuff onto simtel, but it would
be awfully nice if someone could arrange for it.
-geo
---
George D. Hadden, Honeywell Systems and Research Center
PHONE: (612)782-7769
MAIL: 3660 Technology Drive MN65-2100, Minneapolis, MN 55418
ARPA: hadden@src.honeywell.com
UUCP: {umn-cs, ems, bthpyd}!srcsip!haddenjk0@sun.soe.clarkson.edu (Jason Coughlin,,,) (04/19/89)
From article <20609@srcsip.UUCP>, by hadden@ella.SRC.Honeywell.COM (George D. Hadden): > a piece of information: the prolog on simtel (version 1.90n) is > broken. here is one manifestation: if you define append as follows: Don't use pd-prolog. it's crap! here are some major bugs we've found: (a) underline variables are broken. (b) don't name a file the same as any rule or fact (just how long do you think it took us to track down THAT little annoyance?) The worst is that we bought the $200 version and it has the same bugs! Sigh.... Spend your money on Turbo-Prolog instead of wasting your time on sh**ty software that will drive you crazy. Turbo is FAST, cheap, and best of all - it WORKS! Got an flames? Send em to /dev/null. -- Jason Coughlin ( jk0@sun.soe.clarkson.edu , jk0@clutx , or whatever suits your fancy )
hadden@ella.SRC.Honeywell.COM (George D. Hadden) (04/21/89)
>From article <20671@srcsip.UUCP>, by hadden@ella.SRC.Honeywell.COM (George D. Hadden): >> naturally, one well-known problem (some might call this a feature :-) >> of turbo prolog is that it's not prolog. > in what way is turbo prolog NOT prolog? because it forces > you to define the types? personally i consider this a feature > because it makes debugging and tracing easier. > > of course, turbo prolog does have it limitations i DO > conceed. but, it's limitation FAR exceed those of pd-prolog. > also, you can buy turbo- prolog for $40 from an educational > institution. at that price, there just isn't an arguement. > >-- >Jason Coughlin oh, i'd say there's an argument. in fact, i think you've answered your own question. although you might consider the above-mentioned characteristic a feature, that only confirms the claim that turbo prolog is not, in fact, prolog. i'm not sure what it is but it's NOT prolog. please note that this is not a value judgement regarding its worth -- i'm NOT in the prolog fan club. however, what i AM in is a position of teaching prolog to graduate students at the university of minnesota (in the context of an AI Programming Techniques course) and feel an acute responsibility to prepare them for the prolog questions on the qualifying exam. pdprolog, whatever its other shortcomings may be, at least runs the examples in the prolog texts that i use (bratko, c&m). I quote here from the review of Turbo Prolog 2.0 in the October 1988 Byte: Turbo Prolog 2.0 stretches the language even more than did the original Turbo Prolog 1.0 ... Borland's Prologs enforce strong typing of objects and relations (Edinburgh Prolog, the de facto standard, doesn't) and limit the assert and retracdt mechanism to facts alone (Edinburgh supports dynamic modification of facts, predicates that define relations, and rules involving those predicates). ... Like its predecessor, Turbo Prolog 2.0 isn't a traditional Prolog; the language is strongly typed and doesn't let you dynamically assert and retract rules. because of this and other misfeatures of this language, i don't allow my students to use it for their programming assignments. -geo --- George D. Hadden, Honeywell Systems and Research Center PHONE: (612)782-7769 MAIL: 3660 Technology Drive MN65-2100, Minneapolis, MN 55418 ARPA: hadden@src.honeywell.com UUCP: {umn-cs, ems, bthpyd}!srcsip!hadden