The SGML Implementation Guide A Blueprint for SGML Migration

Created: 2020-10-04T10:19:46+00:00

Return to the Index

• Structured documents are important as they allow computers to work with the data.
• SGML was created as a means to bridge markup between working groups.
• Professional typesetting systems once cost 50k$-100k$ USD.
• People made GUI tools to typeset SGML documents and edit DTDs.
• Railroad diagram: a way to visualize parsers.
• Format tagging: when text has style information defined explicitly in the tag.
• Structure tagging: when structural information is in the name of the tag, not implied. ex <h1>, <h2>, instead of <header><header/></header>.
• Content tagging: when tags describe what kind of content is happening.
• A DTD is very important for defining how a particular document is to be read.
• When making a document system; interview users and make their feedback part of the design process. Don't design useless shit nobody uses.
• Parsers have "capabilities" and "features" to identify things like the maximum size of a tag, whether including other documents is allowed.

I gave up around page 252. It goes in to depth on how tags work and are defined. But it's the kind of thing that probably deserves to be in a hypertext reference and not a Zettelkasten card.

DTDs

• ?: optional, zero or one
• +: required, one or more
• *: optional, zero or more
• ,: is followed by
• (): grouping
• |: alternatives
• &: must appear but order is unimportant

Entity

• Maps an entity name to some particular content.
• Apparently this can also import another file.
• Entities also exist so you can type characters that exist in the output typesetting system but not ASCII keyboards. This was also before Unicode.

<!ENTITY onehalf '1/2'>
<!ENTITY chap1 SYSTEM 'chap1.sgm'>

Now when you see &onehalf; it SHOULD be replaced with 1/2.

Element

Defines elements and what is allowed to appear within them.

<!ELEMENT something - 0 (#PCDATA)>

Attribute list

Defines what attributes are attached to tags and whether they are required or not.

<!ATTLIST tagname parameter PARAMETER-NAME #REQIURED>

Document type

Defines the layout of the document.

<!DOCTYPE something [...]>

SGML declaration

Holds "bootstrap" information.

<!SGML ISO 8879:1986...>

Body

• Start tags; ex <foo>
• End tags; ex </foo>
• Entity reference; ex &amp;
• Comment; ex <!-- foo -->
• Processing instruction; ex <?argh ohno>
• Character reference; ex &#45;
• Marked declarations; ex <![IGNORE[foo]]>
• Short reference use; ex <!USEMAP aaaah>
• Link set use; ex <!USELINK #ARGH fuck>

• Start tags may not be required if the grammar says they may be implied.
• End tags may not be required if the tag is marked as empty or the grammar says it is implied.