In June 2005 the Times Literary Supplement carried an article by Martin West containing a much fuller version of Sappho 58 than we'd had before. The extra text had been found in mummy cartonnage at the University of Cologne. This was cause for great excitement, and it didn't take very long at all for a transcription to appear on the Classics-L mailing list.
Unfortunately, there was a small problem — there were two typos in the text, one a spacing issue, and one an interesting metagrammatism — *ἔμαπψε for ἔμαρψε — presumably caused by interference between English "P" and Greek "Ρ" (rho, an /r/ sound). Within hours of appearing on the mailing list, however, the poem — with errors — was all over the internet, on blogs and web pages. There are still some sites that have these errors.
I've recently started loading up Scholiastae.org with some of the major texts, to make it easy for people to drop scholia in without having to deal with some of the initial wikification of a large text (dividing into sections, line number marking, etc.). So I've been spending time looking over already-digitized texts of authors available on the web, which don't appear to be breaking copyright rules and which will be easily converted into wiki format.
When deciding on a text to use for Hesiod's Theogony, I realized that all but a tiny handful of versions of that on the web are from a single source, apparently based on Rzach's old Teubner edition. I know they're from the same source because they all contain the same error in line 268:
αἵ ῥ᾽ ἀνέμων πνοιῇσι καὶ οἰωνοῖς ἕμ᾽ ἕπονται
The underlined word is nonsense, a corruption of elided ἅμα which anticipates the start of the next word — a common enough scribal error. Last time I googled, only four hits show up on that line with the correct ἅμ’, though I've fixed copies I've found on open wikis, and I'll be sending out a few pieces of email. But I do wonder how long the error will persist.