Track full read+write support for the pre-OOXML binary Office formats: Word ([MS-DOC]), Excel ([MS-XLS], BIFF8), and PowerPoint ([MS-PPT]) — all built on the classic OLE/CFB compound-file container ([MS-CFB]).
archive-codec's existing bounded CFB reader already gives each of these a foundation to read the container from — none of the three parse what's inside it yet. A new package covering all three, paralleling how ooxml.js bundles docx/pptx/xlsx into one package for the OOXML family, is the natural home; open to a different split once someone's actually in the spec.
Reopened to track full support as three separate efforts, one per format — see the linked sub-issues.
Original assessment that closed this issue:
Pre-OOXML binary formats ([MS-DOC]/[MS-XLS]/[MS-PPT] over [MS-CFB]). The Open Specifications are authoritative, so this isn't a no-spec problem — but they're enormous (1,000–1,200 pp each for Word and Excel) and remarkably obscure in places. A hand-written codec is categorically harder than anything the family ships: the only binary-format precedent (src/firebird/, src/hsqldb/) is ~1,600–1,900 lines each for far smaller formats.
Permanent out-of-scope under the hand-write-no-library bet. One narrow possible exception worth noting alone: read-only .xls (BIFF8), if ever pursued on its own.
Track full read+write support for the pre-OOXML binary Office formats: Word ([MS-DOC]), Excel ([MS-XLS], BIFF8), and PowerPoint ([MS-PPT]) — all built on the classic OLE/CFB compound-file container ([MS-CFB]).
archive-codec's existing bounded CFB reader already gives each of these a foundation to read the container from — none of the three parse what's inside it yet. A new package covering all three, paralleling howooxml.jsbundles docx/pptx/xlsx into one package for the OOXML family, is the natural home; open to a different split once someone's actually in the spec.Reopened to track full support as three separate efforts, one per format — see the linked sub-issues.
Original assessment that closed this issue:
Pre-OOXML binary formats ([MS-DOC]/[MS-XLS]/[MS-PPT] over [MS-CFB]). The Open Specifications are authoritative, so this isn't a no-spec problem — but they're enormous (1,000–1,200 pp each for Word and Excel) and remarkably obscure in places. A hand-written codec is categorically harder than anything the family ships: the only binary-format precedent (
src/firebird/,src/hsqldb/) is ~1,600–1,900 lines each for far smaller formats.Permanent out-of-scope under the hand-write-no-library bet. One narrow possible exception worth noting alone: read-only
.xls(BIFF8), if ever pursued on its own.