The interest shown in SGML by the federal government has encouraged the development of a number of tools for authoring, revision, and document parsing. Conceivably, these tools can be used to simplify the conversion of text already in electronic form into personal or community hypertext. (249)

Marking a document with SGML is extremely labour intensive. Intelligent front-end tools can alleviate this problem. Indeed, the need for such tools will be responsible for the growth noted by InterConsult, a market research company which has predicted that "[w]orldwide revenues for products related to SGML [...] will more than double by 1995" (Andrews 32).

To be precise, SGML is not a markup standard, but rather a framework for devising such standards. As Williams put it, "[i]t is a standard of semantics to encode documents" (34). Codes may be assigned meanings based on the needs of the author and publisher. Meta-language instructions at the beginning of the document define the symbols in use. SGML tags may be procedural, descriptive, or otherwise.

Barnard et al. have noted another problem with this new standard:

SGML is designed for marking documents that are hierarchical in structure: it assumes that the start- and end-tags that delimit the structural elements of documents will be strictly nested. (266)

Since hypertext documents might be quite non-hierarchical, one of the five approaches they suggest to work around this problem would have to be taken.

Another problem for large documents is that SGML-tagged files must be read in a linear fashion in order to be properly decoded. This results in performance penalties, especially on CD-ROMs. O'Connor has described SGML-B (binary) which allows non-linear access ("Markup, SGML, and Hypertext for Full-Text Databases--Part II" 124).

An important subset of SGML is the Hypertext Markup Language or HTML.