✧✦Catherine✦✧okay so what's going on with office software
the OnlyOffice mess is garden variety xenophobia that people who know better on here are amplifying, that's easy
what the fuck is going on with LibreOffice and Collabora
John Regehr@whitequark has LibreOffice gotten any better recently? god it was awful the last time I tried to use it, and I'll never forget the time one of my students who (with good intentions) moved some of our documents over to it and they were so irrevocably fucked afterwards that we just discarded all of the work he had done
@regehr you'd have to be more specific for me to give you an answer, like what was awful?
@whitequark it just didn't really seem to work, had trouble with basic document formatting and saving and loading and stuff. was a few years ago
Graham Sutherland / Polynomial@regehr @whitequark I similarly found Write to fall short on basic "being a word processor" tasks like reliably formatting & loading/saving when I last used it 4 or 5 years back. Calc seemed less janky at the time.
@regehr @whitequark interop with MS Office was not even minimum viable product quality at the time, too. documents made in Word would be unusably broken when loaded in Write, with stuff off the edge of the page, overlapping text in tables, different fonts and sizes, etc.
if it's gotten to an actually usable standard that would be very good news.
penguin42@gsuberland @regehr @whitequark The file compatibility stuff is really hard; while the MS files XML is publicly specified, it's not very well specified, and there are multiple different embedded formats within those - which may or not be compatible even between MS products. I've fixed a few LO bugs now; last october I did a particularly funny Powerpoint numbered-lists import bug that showed incorrect death stats for a motorbiking club (most were not from biking :-) )
@penguin42 @regehr @whitequark yup. I used to work on this stuff at an old job, around 2012. the OpenXML spec tells you how the documents are structured, but doesn't tell you about special behaviours for specific IDs and indices that are de-facto reserved by MS Office. but that's not really an excuse for how bad it is when it comes to text going off page and tables being broken, cos I got both of those working myself in about 2 days as one of my first Office interop tasks back then.
@gsuberland @regehr @whitequark Right, I agree it does need improving. And yes about the docs not specifying behaviour - compare to say the PDF spec which are very well written; IMHO document format specs should be written as a strict explanation of how to transform the input into a result.
@penguin42 @gsuberland @whitequark I like to imagine the MS devs gritting their teeth and whispering "fuck you" every few minutes as they work on anything remotely supporting interop with their products
@regehr @gsuberland @whitequark That may also happen when the Office 365 online devs deal with compatibility with the desktop or Mac code 🙂
@gsuberland @penguin42 @whitequark I feel like a commercial product dominating the "office" domain was inevitable, since open source would never respond to the whims of large corporate customers the way MS did. but still, this all sucks.
@penguin42 @regehr @whitequark and that was in Delphi without an XML library. just rawdogged the strings. horrible but it worked.
for Excel interop I ended up writing a Delphi-to-.NET marshalling interop layer so I could load the ClosedXML library, which absolutely slaps and takes a lot of the pain away.
@gsuberland @regehr @whitequark How much did you actually implement? I mean it's not that hard to read/generate files with a handful of features; e.g. parsing a report generated by a specific company/word version, or generating a report with a handful of features - the tricky bit is implementing all of it; they say people only use a few % of a programs features, alas everyone uses a different few %.
@penguin42 @regehr @whitequark oh I know it's hard to implement all of it, but I had all the different font styles, sizes, etc. at the paragraph (named style) and formatted section levels, bulleted lists, links, tables with border styles, auto/explicit sizing, merged cells, images, and some other stuff, and supporting load/save. the biggest hurdle was just initially figuring out the reserved ranges where you couldn't set styles, once I got that done the rest came quick.