Apologies if this is answered elsewhere. I was wondering if anyone knows a way of converting the mishmash of symbols used to represent Greek in older textkit posts into something readable. E.G.
[face=spionic]h0ni/d’, w} Turoi=, fili/a nea~nij,
nu~n pe/feukaj, eu0pate/reia mo/sxe
kalli/morfe/ t’ Ai0oli/dou tura/nnou,
dw&mata traxh~.
I may write a script to fix it in the database. However, until that time, here is a cheat key: http://www.drshirley.org/fonts/SPIonic.html
And a converter: https://noahfrederick.com/spionic-unicode-converter/
ἠνίδ’, ὦ Τυροῖ, φιλία νεᾶνις,
νῦν πέφευκας, εὐπατέρεια μόσχε
καλλίμορφέ τ’ Αἰολίδου τυράννου,
δώματα τραχῆ.
Noah Frederick was kind enough to share his converter code with me when I asked him about it, and he has in fact opensourced it here: https://github.com/noahfrederick/spionic I’m really thankful to him for this, and it has made a big difference for being able to fix Textkit’s decade-old SPIonic posts.
I’ve used the above, along with a fair amount of glue code, to convert over all the face=spionic text in the database. Some of it was actually beta code masquerading as SPIonic, but I wrote some code to differentiate. However, that code was not perfect, and there are some posts that still have bad characters, though they are readable.
The most annoying defect to my change is that we lost line breaks in the SPIonic Greek text for some posts. The 2005 post about Latin and Greek odes that started this thread is an example. The Greek linebreaks are gone, though the text is now readable.
two odes, two translations, two epigrams
(Aside: The above is really good.)
εὖγε 
Thank you!