Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!ut-emx!tivoli!lark From: lark@tivoli.UUCP (Lar Kaufman) Newsgroups: comp.text.sgml Subject: Re: looking for more information Message-ID: <200@tivoli.UUCP> Date: 19 Nov 90 21:50:19 GMT References: Reply-To: lark@tivoli.UUCP (Lar Kaufman) Organization: Tivoli Systems Inc., Austin, TX Lines: 111 In article dms@aix03.aix.rpi.edu (david m schwartz) writes: >Hello -- I became aware of SGML this fall. The folks who market SGML had an >exhibition booth at EP '90 (Electronic Publishing Conference held this year at > ... SGML is not a proprietary product. It is an ISO standard. > >1. SGML has been described as a context-free markup. In a recent posting I >think I read someone refering to TeX as a context-specific markup. Could >someone enlarge on this thread for me? That would be a big task. Continue. > >2. Latex (I believe I also read in a recent article) is a more suitable >SGML environment than is TeX. Is this because Latex is a macro-based >markup while Tex is more a control-word type markup. In other words, a markup >that describes what the text looks like vs. what document element the text is >being assigned to? Would this analagy hold also, for say, SCRIPT vs. GML? >GML is a fore-runner of Bookmaster. An intermediate step was ISIL (which never >did make it past internal use at IBM) but there are some very significant >bloodlines here and if someone can frame their answers to my questions around >these IBM products, it will greatly add to my understanding. In fact, I >read someone's comment that Bookmaster is a suitable SGML environment. I can oblige you here. Script is similar conceptually to troff, TeX, Scribe, and the like. GML is a structured language that is interpreted and converted to script, which in turn is converted to device driver instructuctions for printing. ISIL is just an extended set of GML (Doesn't ISIL mean IBM Structured Information Language?) and Bookmaster is a more commercialized packaging of ISIL. Bookmaster is complemented by a package called BookManager, which is intended to provided formatting and presentation of online documentation. GML was written by Charles Goldfarb. ISIL and BookMaster were extensions for IBM's not always apparent motivations (but you can be sure that money is a factor). Unfortunately, ISIL and Book- Master/Manager took directions that led them astray from the "purity" of GML and away from the direction pointed to by SGML. (SGML was also a Charles Goldfarb conception.) However, IBM (or elements within IBM) have come to recognize that SGML is A Good Thing. Therefore, IBM's latest version of BookMaster/Manager can parse SGML as well as its ISIL-flavored code. That is, if I understand correctly, BookMaster 3.0 can parse SGML. My disclaimer: I have contracted to IBM in the past and used these products, but I am not privy to their plans or reasoning. I have communicated with Charles Goldfarb, but only seeking information about SGML and not about his work for and with IBM. > >4. Is it likely that we will see SGML front-ends on future word-processors? But of course. Sort of. Word Processor and Typesetting packages will happily convert SGML into their proprietary formats. A few farsighted companies will provide the reverse conversion (a more difficult task). A few clever companies will use SGML as their native format. > >5. As I currently understand it, SGML is a two-step process: > > a) define a structure for the document > b) write the document > c) run the document through the SGML software (?) and see if the > document adheres to the structure. > >three, three steps, as I understand it, SGML is a three step process >no one expects the Spanish Inquisition, shades of Monty Python :) Actually, most users will not mess with structure definitions. They will use predeveloped Document Type Descriptions. Formatting gurus will do the document definition stuff. The point is, the average writer shouldn't have to (be allowed to) mess with the document format. >6. I assume it is OK for SGML documents to be broken into chapter files? Yeah. >... I hope that these questions make some sense. They do, but they suggest that you might want to "read up" on SGML. > >FYI, my interest in all this comes from the fact that I perform technical >writing services for a company that develops a text-retrieval program. The >growing popularity of SGML is making the developers scurry to permit not only >the input of an SGML document, but the retrieval of said document by >permitting the user to specify search parameters that are unique to the >hierarchical structure of an SGML document. Yessss, any company doing text-retrieval must be very interested in SGML. I don't know that I think of an SGML document as necessarily being hierarchical... I would suggest further readings. Unfortunately, the publishing software companies have been very slow in moving towards SGML, in spite of early endorsement by the American Association of Publishers. The real impetus is the U.S. Government's CALS initiative, and the potential for amazing online documentation capabilities (including text retrieval). I have "SGML: An Author's Guide" by Martin Bryan (Addison Wesley). It is an OK starting place, abeit of narrow focus. I am waiting for "The SGML Handbook" by Charles Goldfarb, which I already expected to be out by now... maybe I'd better start hounding my favorite bookstore... The SGML Handbook should incorporate the ISO 8879 specification, so you can safely skip getting that document. If you are involved with IBM products, you can examine "Solutions for CALS Technical Publishing" (IBM document GC34-5153). I cannot help further, because I am still in early stages of research myself. (Need to spend a few days in a good library, going through the periodical literature and stuff). Sorry for chopping up your posting and ignoring part of it, but time flies... -lar -- --------- TIVOLI Systems, Inc. Lar Kaufman 512-454-3301 (voice) 512-329-2455 4503 Sinclair Avenue (fax) 512-329-2755 Austin, Texas 78756 USA (e) lark@tivoli.com