Summary
Two of the 38 files in cd/Official/ declare themselves experimental:
| File |
<CDStatus> |
<CDReviewDate> |
scscp1.ocd |
experimental |
2017-12-31 |
scscp2.ocd |
experimental |
2017-12-31 |
The other 36 declare official.
Why it matters
A consumer deciding which dictionaries to trust has two signals that disagree,
and no way to tell which one you consider authoritative. Treating the directory
name as the answer — which it rather invites — silently pulls in experimental
content; treating <CDStatus> as the answer means the directory layout is
misleading.
I hit this harvesting the directory as "the official set" for a conformance
suite, and only noticed because the three malformed examples I reported in #44
turned out to be in one of these two files, which changed how seriously I took
them.
Possible resolutions
- if the directory is normative: move
scscp1.ocd and scscp2.ocd to
cd/experimental/;
- if
<CDStatus> is normative: a line in the README saying so would save the
next person the same question.
Also worth noting that both carry a CDReviewDate of 2017-12-31, which by the
standard's own convention marks them as due for review.
I have no stake in which way it goes — I would just like to be able to tell.
Disclosure: this report was drafted with AI assistance. The findings are not
generated text — they come from running an independent OpenMath 2.0
implementation locally over a clone of this repository. The file count, the two
<CDStatus> values and the review dates were read from the files by that run and
re-checked before posting, so they can be reproduced rather than taken on trust.
Summary
Two of the 38 files in
cd/Official/declare themselves experimental:<CDStatus><CDReviewDate>scscp1.ocdexperimentalscscp2.ocdexperimentalThe other 36 declare
official.Why it matters
A consumer deciding which dictionaries to trust has two signals that disagree,
and no way to tell which one you consider authoritative. Treating the directory
name as the answer — which it rather invites — silently pulls in experimental
content; treating
<CDStatus>as the answer means the directory layout ismisleading.
I hit this harvesting the directory as "the official set" for a conformance
suite, and only noticed because the three malformed examples I reported in #44
turned out to be in one of these two files, which changed how seriously I took
them.
Possible resolutions
scscp1.ocdandscscp2.ocdtocd/experimental/;<CDStatus>is normative: a line in the README saying so would save thenext person the same question.
Also worth noting that both carry a
CDReviewDateof 2017-12-31, which by thestandard's own convention marks them as due for review.
I have no stake in which way it goes — I would just like to be able to tell.
Disclosure: this report was drafted with AI assistance. The findings are not
generated text — they come from running an independent OpenMath 2.0
implementation locally over a clone of this repository. The file count, the two
<CDStatus>values and the review dates were read from the files by that run andre-checked before posting, so they can be reproduced rather than taken on trust.