Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!tut.cis.ohio-state.edu!rutgers!mit-eddie!husc6!bunny!slzr From: slzr@bunny.UUCP (Suzanne Sluizer) Newsgroups: comp.lang.prolog Subject: Re: Arity Prolog / Expert Development Package Keywords: Prolog, expert systems, development environment Message-ID: <6765@bunny.UUCP> Date: 7 Apr 89 18:23:26 GMT References: <47016@linus.UUCP> Reply-To: slzr@bunny.UUCP (Suzanne Sluizer) Followup-To: comp.lang.prolog Distribution: comp Organization: GTE Laboratories Lines: 61 In article <47016@linus.UUCP> wdh@linus.UUCP (W. Dale Hall) writes: >Does anyone out there in net.land have any experience with the >products of Arity Corporation? I am considering obtaining their >"Combination Package" which has Prolog (compiler and interpreter), an >Expert Systems Development Package (presumably a shell and some >built-in functions to make development a tolerable experience), an SQL >Development Package (for database manipulation), and some form of >Prolog toolbox. I have been using Arity/Prolog for over 2 years now. (I use the compiler and interpreter. I also have limited experience with the toolbox, but none at all with the expert system package.) We have developed an executable specification/modeling tool, which consists of more than 100 pages of Prolog. It runs under C-Prolog, Quintus Prolog and Arity/Prolog. In general, the code that runs under one Prolog runs under the others (there is less than one page of code difference among the three). Normally, I do most of my development using Arity/Prolog. I tried to put up our modeling tool under ALS Prolog (on the PC, with the idea of eventually transferring it to the MAC). However, I discovered many problems with ALS and it was unable to handle a program of the size of our tool. Arity/Prolog Versions 5.0 and 5.0x (and presumably 5.1) handle it easily. I started with Version 4 and upgraded to Version 5.0x last summer. Version 4 was slow, and didn't allow you to control much in the way of your environment (e.g. heap sizes). Howver, if you wrote "kosher" Prolog, it worked well enough. Versions 5.0 (and higher) allow much more control, which is important if you write large Prolog programs. Versions 5.0 and 5.0x were incredibly buggy when I got them. It took about 1 1/2 days to transfer our system from Version 4 to Version 5.0x. Many problems were only resolved by calling Arity/Prolog customer support and asking for help. (For example, "reconsult" did not work properly and "listing" did not work properly.) However, once these initial problems were resolved (an interim Version was downloaded from the Arity customer bulletin board), I had no further problems. Other people here found additional problems with Version 5.0x (the interface to C didn't work well and there was a problems with definite clause grammars). However, Versions 5.0 and 5.0x ran like "a bat out of hell". It was much faster, for example, than interpreted C-Prolog on a Vax 750 or a Microvax (and this was just interpreted Prolog). However, the new Version 5.1 seems to have solved all these problems, according to my colleagues at GTE Labs. I have not been able to try out 5.1 as I just received it a week or two ago, and my PC died recently and is undergoing major repairs. I recommend Arity/Prolog. They have a dialect of Prolog that is similar, but not identical, to Clocksin & Mellish. It is based on Dec-10 Prolog. I have also found the customer service people at Arity to be quite helpful. However, the Arity sense of time seems to be different than mine. When I ran into problems last summer with 5.0 and 5.0x, I was told that 5.1 would be shipped "in a few weeks". I just finally received it a couple of weeks ago. -- Suzanne Sluizer CSNET: slzr%gte.com@RELAY.CS.NET GTE Laboratories UUCP: ...!harvard!bunny!slzr 617-466-2882 "Truth is a pathless land." -- Krishnamurti