This afternoon I made two decisions about CaseRelief that I think everything else is going to be built on. One was how your email gets in. The other was what never gets taken out.
Gmail connects with an app password. It's a 16 character code you make in your own Google account and paste into CaseRelief. You need 2-step verification turned on first, and it takes a few minutes. iCloud works the same way. For Outlook the plan is a proper Microsoft sign-in, since Microsoft is getting rid of app passwords anyway.
Then I thought, why not just use email people have already downloaded? If you've got your email in some form, the app should take it. A file doesn't need a login or a password or a permission screen, and the app never touches your account at all.
CaseRelief now takes .mbox files, which is what Google Takeout gives you and what Thunderbird and Apple Mail can export. It takes .eml files, one at a time or a whole folder of them. And it takes the .zip that Takeout actually hands you, so you can drop that straight in without unzipping anything. Everything after that works the same wherever the email came from.
I also redid the file it puts out. Who else got copied is often exactly what you need to show, so CC is on every email, along with From, To, date and subject. Every email gets a number (#0001 and up, with more digits once you pass 10,000), so you can point someone to "email 47" and they'll find it. A reply links back to the email it's answering if that one's in the file too. My one rule was that it always has to be readable by a person, not only by software.
Then I ran it on my own mailbox and read through what came out. Some threads were quoted inside themselves over and over, more than a dozen levels deep, every reply carrying everything before it. The file got long.
The obvious fix was to collapse the repeated quotes or swap them for a pointer to the earlier email. I went back and forth on it and decided no.
The quoted text in a reply is sometimes the only copy you have of what the other person said. If they wrote to someone else and you only got added later, their words might only exist in your mailbox as a quote under somebody's reply. Strip the quotes and that's gone.
Representing myself, I've also seen how much context changes things. A short, sharp reply can look bad on its own and completely reasonable once you see what it was answering. And people's words should stay the way they wrote them. If I paraphrase someone, it's my version of what they said.
Spotting duplicates is harder than it sounds too. People edit what they quote, and mail programs reformat it. Any rule smart enough to strip out repeats is eventually going to strip something that wasn't a repeat.
So CaseRelief keeps reply chains exactly as they were sent.
– M