3 ms·
probably a legacy from stardivision, the german makers of the original staroffice, which became openoffice.org, which in turn became libreoffice. it's origins g
by tech-no-logical 12y ago
probably a legacy from stardivision, the german makers of the original staroffice, which became openoffice.org, which in turn became libreoffice. it's origins go back to when the z80 was a 'current' processor :)
- davidgerard 12y agoThis is what is interesting about LibreOffice. It's a thirty-year-old C++ codebase. Computers now have 100,000 times the memory and 1,000,000 times the CPU! And this codebase has had mediocre commercial coders working on it out of public view and following the hairpin turns of marketing. And commenting in German, because they were substantially in the StarDivision office. It's very messy. But it's an actual product that, for all that, basically works. And millions of real people do real work with it every day. I find Michael Meeks' tech reports really interesting, because they show how on earth a small band of very high-calibre software engineers can possibly maintain a several-million-line codebase with far too few people. The answer is automation, directing slogging through. Unit tests, automatic code checkers, Coverity, clang plugins, assorted torture tests ... It's a useful example of how too few people can effectively maintain and develop a huge codebase. This is knowledge the world will need for the future, as software continues to expand as absolutely large as it can.
- renox 12y agoWell, the OTHER useful example is STEPS project of the VPRI institute which claims that you could build the equivalent of our modern desktop software which much, much less code. The big issue with this project is that AFAIK no code is available.. But it could be an interesting way: maintain an existing enormous codebase and build from scratch in parallel another one, then meet at the middle.