Article View: comp.text.sgml
Article #192GML or BookMaster to SGML
From: "Dr. "Eliot Kimber" Macro"
Date: Sun, 16 Aug 1992 16:41
Date: Sun, 16 Aug 1992 16:41
78 lines
4689 bytes
4689 bytes
References: <1992Aug11.051321.12743@a.cs.okstate.edu> Rick Matzen writes: > Does anyone know of an available translator for either GML to SGML > or Bookmaster to SGML: for (GML/Bookmaster design specs to DTD's) and/or > GML/Bookmaster documents to SGML documents (instances)? Part of the IBM SGML Translator product is a GML2SGML converter designed to convert GML Starter Set documents to SGML Starter Set documents. This will also work against BookMaster documents (it just changes colon/period to angle brackets, for the most part), but you will still need a DTD that describes BookMaster, which doesn't exist in any form officially supported by IBM (mostly because BookMaster as it exists today represents some serious problems when trying to represent it in SGML, and thus, the BookMaster language itself needs to be re-defined to be done "right"). Your best bet is to contact your local IBM representative and have them contact Publishing Marketing in Boulder. They may be able to help you. Because we face it ourselves internally, we recognize the need for migration aids and assistance for moving from GML Starter Set and BookMaster to SGML languages. Publishing Marketing has available to it tools it can provide "as is" to customers to solve specific problems like these. Your IBM rep can find out what's available at any given time. In addition, you can have your rep contact me as I may be able to provide additional assistance. Unfortunately, I will be on vacation and travelling on business the next three weeks, but I'll try to check my mail when I can. If you have existing BookMaster documents that don't use the full range of BookMaster tags and are pretty well coded (in other words, you followed the precepts in the BookMaster User's Guide closely and didn't do anything you shouldn't have), you can probably cobble together a DTD based on the GDOC DTD that comes with the Translator that represents your local documents. The things you typically have to add that are not in the SGML Starter Set are: o Labeled boxes (labeled boxes can contain just about anything) o Parameter lists and glossary lists o CGRAPHICs and SCREENs o PV, PK, MV, RK, CIT, HEX, BIN, CHAR phrases o TDEF, which is RDEF in the Starter Set o PROLOG and related front matter elements. The PROLOG content model turns out to be a real bear. Try not to use large & groups. I've found it more profitable to impose a required order on "code zero or once" elements rather than try to use & connectors. The elements DOCPROF, TITLE, DATE, DOCNUM, PRTLOC fit this requirement. All other PROLOG elements are "code zero or more" and can be in a large OR group. o Major document sections (BODY, FRONTM, BACKM, APPENDIX). Again, impose an order even though BookMaster doesn't require one, otherwise you get huge AND groups that most DTD compilers will choke on. If you are constrained to the reference concrete syntax declaration, as you are with the current release of the SGML Translator, you will hit the 96-tag limit for some content models if you try to support all BookMaster tags. You'll have to make choices. Note too that any DTD that supports what most BookMaster documents use will blow the RCS capacities by a factor of 2 or 3 (GRPCAP upwards of 70K instead of 35K). If you can define your own declaration, I recommend a name length of 8 and a TOTALCAP of 100 or 200K, with a large GRPCAP. Because of the large number of different GIs and the ubiquitousness of those GIs across content models, GRPCAP seems to be the capacity that gets hit the hardest. Note that the BookMaster Release 3.0/4.0 language is probably a worst-case when expressed in SGML. It is too big and too general, and probably represents the biggest single DTD anyone will ever define. A more rational approach to converting BookMaster to SGML will involve simplifying the language and modularizing it to enable to the creation of smaller DTDs from a common base. Eliot Kimber Internet: drmacro@ralvm13.vnet.ibm.com Dept E14/B500 IBMMAIL: USIB2DK9@IBMMAIL Network Programs Information Development Phone: 1-919-543-7091 IBM Corporation Research Triangle Park, NC 27709
Technical Details
Message-ID:
<9208161631.AA04071@ucbvax.Berkeley.EDU>Path:
arc1-ca.pugleaf.net!sparky!uunet!seismo!darwin.sura.net!wupost!zaphod.mps.ohio-state.edu!cis.ohio-state.edu!ucbvax!RALVM13.VNET.IBM.COM!DRMACROShow Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!seismo!darwin.sura.net!wupost!zaphod.mps.ohio-state.edu!cis.ohio-state.edu!ucbvax!RALVM13.VNET.IBM.COM!DRMACRO
From: DRMACRO@RALVM13.VNET.IBM.COM ("Dr. "Eliot Kimber" Macro")
Newsgroups: comp.text.sgml
Subject: GML or BookMaster to SGML
Message-ID: <9208161631.AA04071@ucbvax.Berkeley.EDU>
Date: 16 Aug 92 16:41:57 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Lines: 78