enrich-from-pages-you-do-not-control
The procedure as a graph, on its own page — zoom it, print it, send somebody the address. Back to the record for the steps, the trust panel and the provenance.
Procedure
The same procedure, read in order
- start start. Goes next to s1
- check Fetch one page, read its status code. Goes on_fail to s10; next to s2
- decision What did the fetch actually get?. Goes if refused to s3; if needs_browser to s4; if gone to s9; if default to s5
- action Record the refusal as a refusal, RW (reversible-write). Goes on_fail to s10; next to s9
- action Queue it for a browser, after the cheap pass, RO (read-only). Goes on_fail to s10; next to s9
- decision Does the page carry structured markup?. Goes if prose_only to s6; if default to s7
- action Write a rule for this family, not for all, RO (read-only). Goes on_fail to s10; next to s7
- decision Did this field ever arrive?. Goes if never to s8; if default to s9
- action Stop promising the field, or take it from text, RW (reversible-write). Goes on_fail to s10; next to s9
- action Write what was read, and leave the rest empty, RW (reversible-write), requires approval. Goes on_fail to s10; next to __end_success
- escalate Stop before the batch runs. Terminal
- end success. Terminal