Path: utzoo!utgpu!water!watmath!clyde!rutgers!sri-spam!ames!elroy!mahendo!jplgodo!wlbr!scgvaxd!ashtate!dbase!drc From: drc@dbase.UUCP (Dennis Cohen) Newsgroups: comp.sys.mac Subject: Re: MacUser compiler comparison Summary: Pascal Comparisons Message-ID: <296@dbase.UUCP> Date: 29 Jan 88 15:20:09 GMT References: <10870@shemp.UCLA.EDU> Organization: Ashton Tate Development Center Glendale Cal. Lines: 90 Before I get into my comments on the following, it should be noted that I frequently review compilers for a couple of competing publications (MacWorld and Macintosh Today -- MacUser never bothered to respond to any queries about Author/Reviewer guidelines {much less anything else I asked}). In article <10870@shemp.UCLA.EDU>, khayo@sonia.cs.ucla.edu (Erazm J. Behr) writes: > Hi; I have a question to those who are familiar with both > Turbo & LS Pascal compilers. In the latest MacUser I saw > an article comparing various Mac compilers (not only Pascal). > As an owner of TP I quickly looked at a table where points were > assigned in several categories, and saw that TP scored pretty > low in all, which is of course clear heresy :-) I know it's > far from being perfect, but: The first thing you should be aware of is that ratings such as those in the table, lacking any examples of deficiencies, are purely subjective as they don't specify what is required for the various ratings in these categories. My personal favorite as "best" of the compilers is MPW, a preference not all will share, because it suits _my_ needs best. > it got 3 or 4 out of 5 for "Toolbox access" and something like > 3 for "standalone applic."; according to the manual *full* access > to the toolbox is provided (and is easy, in my opinion) and > creating standalone applications is a snap (only the dumb loader > makes small ones big and big ones a little bigger :); what gives? > What does LSC (perfect scores in all categories) have that TP > doesn't in this regard? The version of TP used for the article was 1.0 and it didn't support anything from Volume V of Inside Mac, which should have dropped it to a IV, no lower. Creating a standalone application is probably easier in Turbo than in LSP for the following reasons: 1) You don't have to leave the environment to edit your resource file 2) You can set the Creator and your bundle bit with Turbo and cannot with LSP. The major thing which is easier in LSP than any of the others is debugging, but even that isn't enough to make me put up with their editor unless I am really desparate (Very Personal and Subjective Opinion :=}). > on the other hand, the only category in which TP beat TML is > "listing"; from a quick glance at the accomp. text I gathered this > was about generating x-references etc. If TML's ability to do this > is *worse* than TP's, then I feel *really* sorry for those poor > guys who use TML (I mean, TP's listing is "pure, unadulterated" text > of the program!). They must not have seen any of the _Public Domain_ tools that Ken Butler and I have provided, such as PasMat and PasRef for both TML and Turbo. These tools are available on CompuServe and I submitted them to the moderator of binaries here about 6 months ago (though I've never seen them come past). Come to think of it, I've never seen anything that I submitted to a moderator come by on the net. Both PasMat and PasRef are standalone application equivalents of the MPW tools by the same names. They are public domain because their original versions came into existence on Univac 1100 Series mainframes in the late 70s and were created by government contractors and employees on the public dime. The versions from Apple and from me for the Mac are significantly enhanced deriv- ative works. BTW, TML's listing is essentially identical to the listing from Turbo (straight editor printout). The only reason I can see for LSP getting a good grade here is that it bold-faces reserved words (personally, I would take points away for that, but that too is a personal preference). > So, have the scores been assigned randomly, or did the editors > use substances, or what ? Eric As stated above, if they had criteria for their grading, it was not stated explicitly at any point. I personally took the whole article as a piece of fairly useless fluff. In conclusion, I own each of the four compilers discussed, as well as four C compilers for the Mac, three Modula-2 compilers, and two assemblers. I use all of them at least occasionally (with the exception of TDI Modula-2, which bombs too frequently). In general, I don't believe that there is anything better than MPW for the professional developer, though LSC is neck and neck with MPW C. Among the Pascals any of TML, Turbo, or LSP could be right, depending on the way you like to do things. I don't use LSP much anymore because it doesn't "feel right" to me, doesn't work with MultiFinder and even the Mac II-compatible version has some problems when you run in color. The available Volume V stuff is also wrong in many places. Turbo 1.1 is an excellent product and has full Volume V access. The two things which I would add to it would be an integrated resource compiler and source level debugging. TML 2.5 is a solid product which is fully Mac II and MultiFinder compatible, with full Volume V interfaces. As a matter of fact, with MultiFinder and enough memory (about 2.5 MB or more), it's a pretty good integrated product. It's not as pretty as the others, but its linker is much better and it has Object Pascal extensions (although this area is buggy). Dennis Cohen Ashton-Tate Glendale Development Center dBASE Mac Development Team -------------------------- Disclaimer: Virtually everything above is, as stated, my personal opinion and has no relation to any corporate position.