RCS file: RCS/dvistd0.tex,v Working file: dvistd0.tex head: 5.1 branch: locks: strict schrod: 5.1 access list: symbolic names: draft_0-05a: 5.1 draft_0-04c: 4.4 draft_0-04b: 4.3 draft_0-04a: 4.2 draft_0-04: 4.1 draft_0-03: 3.1 draft#002: 2.1 draft#1: 1.1 comment leader: "% " keyword substitution: kv total revisions: 13; selected revisions: 13 description: DVI driver standard, level 0 ---------------------------- revision 5.1 locked by: schrod; date: 1991/08/06 15:12:07; author: schrod; state: Released; lines: +481 -213 draft 0.05a A LOT OF CHANGES; DRIV-L POSTINGS OF THE LAST YEAR COLLECTED AND PARTLY INCORPORATED. EDITORIAL CHANGES: Did not use style option mf any more. This style option needed the NFSS, which is not in use at all sites. The draft must be TeXable with a `standard' LaTeX system. The `author' of the draft is now the TUG DVI Driver Standards Committee. Added distinction between standard text, explication, and rationale. The standard text is just the `pure,' definitive, short statement to be made. An explication pinpoints important consequences or implications of the standard text. Both are published in one document and are already finished. The rationale is an additional document describing why this standard text was chosen and names other possibilities discussed, but not included. The reason for the possibilities which were not included are usually outlined. The rationale is not written yet. All three parts will be within one \TeX{} source, although the rationale will be published as a separate document. (This conforms more to the standard documents I've read up to now, this structure may change again in the future.) Completed replacement of `should' with `shall' and `must.' (`should' is no good term in a standard...) There's still one `should': In the section on specials, on the definition when warnings should be issued. But this section needs clarification anyhow. Completed replacement of `DVI driver' with `DVI processor.' (At least on those occasions where not explicitely a device-driving program is meant.) Exchanged item `Magnification number' and `Resolution number' (now first) since the latter should be the canonical form by now. CHANGES OF CONTENTS: Abstract slightly rewritten, it conforms now more to usual conventions (see, e.g., Mark Wegman's paper in SIGPLAN, Vol. 21, No. 5). Commented out the trip test statement in the abstract. A trip test is not mentioned in the report. If it is mentioned it will be added back. Enhanced the introductionary paragraph to define a few terms (render, etc.) used later on. The term `DVI font' was defined nowhere and is therefore discarded. Just `font' is enough, since it concerns all fonts a DVI processor must handle. The same holds for `DVI character size,' it's renamed to `character size.' A paragraph was added to the explication of the section `character size' which states explicitely that the `character size' is the size of the glyph, not the TFM bounding box. This explication was demanded by Pierre MacKay on driv-l and generally aggreed on. The term `non-unit aspect ratio' does not exist. Renamed to `aspect ratio unequal to one.' Added explication which points out that no rule is set if the width or the height is less than 0pt. Added an explication that the maxdrift algorithm does not solve all problems with positioning. Mentioned a few problems. Added escape-clause (realization of functionality perhaps not possible due to device contraints) to the section `Objects off the page' (formerly named `Objects off of the page'). It was forgotten when accepting ammendment 06. Defined the term `special.' Says explicitely that level 0 does not define the meaning of any special. In section `Configuration' changed `without having to recompile the processor' to `without having to recompile or relink the processor.' Added explication what `etc.' in this paragraph means (`make as many things configurable as possible'). But this explication should be changed anyhow... In section `Minimum set of magnifications,' added explication which defines the term `magstep.' Moved the encouragement to support all possible magnifications to the explication, it's implicitely contained in the standards text. ACKNOWLEDGEMENTS: Small changes for correction of spelling errors and inserting better phrases (contributed by Nelson Beebe, Karl Berry, Friedrich Haubensak, Berthold Horn, Pierre MacKay, Doug McDonald, Arthur Ogawa, Greg(?) Platt, Liam Quin, and Thomas Reid). ---------------------------- revision 4.4 date: 1991/03/25 00:00:00; author: hosek; state: Released; lines: +20 -9 draft 0.04c Merged in ammendments 09 and 10. Ammendment topics: 09: definition of round() (S2.6.2) [introduces and defines pixel_round()] 10: change of max_drift correction (S2.6.2) [move only one pixel] Corrects some typographic errors of draft 0.04b, but again, this document is not TeXable. ---------------------------- revision 4.3 date: 1991/01/24 00:00:00; author: hosek; state: Released; lines: +24 -16 branches: 4.3.1; draft 0.04b Typos sent in by me and others are corrected. But still some typographic errors remain... ---------------------------- revision 4.2 date: 1990/11/13 00:00:00; author: hosek; state: Released; lines: +59 -29 branches: 4.2.1; draft 0.04a Merged in a lot, but not all of my changes sent to driv-l. Inserted stuff which was sent as ammendment 11, making the ammendment obsolete in this way. This covers the limits for DVI units, specification of movement, unconditional warning if fonts are missing. But still some typographic errors remain. This draft was even not TeXable when it arrived... ---------------------------- revision 4.1 date: 1990/10/06 00:00:00; author: hosek; state: Released; lines: +288 -110 draft 0.04 Merged in a lot of minor changes, e.g., render instead of print, must instead of should, etc. Added difference between standard text and explication (for a future rationale document). Merged in ammendments 01, 05--08. Ammendments topics: 01: Change margin of error tolerance (S4.3.2) [font mag epsilon: 0.2% instead of 2%] 05: Change to wording of minimum stack depth (S2.5) [allow dynamic stacks] 06: DVI limits only for devices which may support it (S2) [some limits were not *possible* to print] 07: Change to definition of limits for maxdrift (App E, now S2.6.2) [now TFM independent, since driver may not read TFM files] 08: Change to maxdrift value (App E, now S2.6.2) [0 at 100dpi] Moved maxdrift algorithm from Appendix E into the core text. ---------------------------- revision 3.1 date: 1990/07/19 00:00:00; author: hosek; state: Released; lines: +65 -54 branches: 3.1.1; draft 0.03 Official draft enumeration changed from just a number to x.y with x as a designation for the standard part and y as a designation of the version. Adopted the notion of ``tiers'' for the future structure of other standard levels. Added appendices with max drift algorithm and file format specifications. (CAVEAT: The max drift appendix is wrong!) ---------------------------- revision 2.1 date: 1990/06/23 00:00:00; author: hosek; state: Released; lines: +267 -112 draft #002 This draft was the first one which was written with LaTeX markup, and in complete sentences, not just phrases. ---------------------------- revision 1.1 date: 1990/02/26 00:00:00; author: hosek; state: Experimental; Initial revision ---------------------------- revision 3.1.1.1 date: 1990/08/02 00:00:00; author: schrod; state: Experimental; lines: +17 -4 Covers ammendments 01 and 05 (02--04 failed). These ammendments were merged in by me. Ammendment topics: 01: Change margin of error tolerance (S4.3.2) [font mag epsilon: 0.2% instead of 2%] 05: Change to wording of minimum stack depth (S2.5) [allow dynamic stacks] ---------------------------- revision 4.2.1.1 date: 1991/01/29 16:15:08; author: schrod; state: Experimental; lines: +23 -16 Corrects the typographic errors of draft 0.04a. Will hopefully be 0.04b. ---------------------------- revision 4.3.1.1 date: 1991/01/29 17:22:58; author: schrod; state: Experimental; lines: +14 -7 branches: 4.3.1.1.1; 4.3.1.1.2; Corrects the typographic errors of draft 0.04b. Will hopefully be 0.04c. ---------------------------- revision 4.3.1.1.2.1 date: 1991/01/29 17:48:59; author: schrod; state: Experimental; lines: +13 -8 draft 0.04b with correction of typographic errors and ammendment 09 Ammendment topic: 09: definition of round() in section 2.6.2 Introduces and defines pixel_round(). ---------------------------- revision 4.3.1.1.1.1 date: 1991/01/29 17:34:35; author: schrod; state: Experimental; lines: +9 -4 draft 0.04b with correction of typographic errors and ammendment 10 Ammendment topic: 10: change of max_drift correction Do not move max_drift pixels if the distance is larger, but just one pixel. This keeps the movements within one word smaller. Note: This is the same as in DVItype. This was an long outstanding error in the draft standard. Although it was pointed out very often an ammendment was needed to correct it. =============================================================================