Hi folks, just continuing the chat here from the other thread about the retrosoftware.co.uk WikiMedia dump repo we added to Github.
I managed to find my hacky python script that did some of the conversions and have added it to the repo for reference:
https://github.com/stardot/retrosoftwar ... -converter
Its over 5 years old, so I cant remember much of how it worked but I do remember a lot of html tags -> markdown edge cases, handling tables and such.
There's a sample of some the output here: https://github.com/stardot/retrosoftwar ... erter/wiki
I've no idea if what's in there is a full snapshot of what was on the site when it went offline, but there's a very hacky github pages hosting that markdown here now: https://stardot.github.io/retrosoftware.co.uk/
retrosoftware.co.uk archive
Re: retrosoftware.co.uk archive
Thanks for sharing this! I see that it leans on Pandoc for the format conversion.simonm wrote: ↑Thu Jan 19, 2023 12:42 am Hi folks, just continuing the chat here from the other thread about the retrosoftware.co.uk WikiMedia dump repo we added to Github.
I managed to find my hacky python script that did some of the conversions and have added it to the repo for reference:
https://github.com/stardot/retrosoftwar ... -converter
As I noted in the other thread, I can imagine that targeting Markdown is infuriating given the limitations of that format. One MediaWiki syntax feature that is likely to be problematic is its support for tables, where a convenient multi-line construct is available. Other markup languages that are keeping up the pretense of resembling plain text tend to only allow single lines of source text per table row, with some even insisting that all the separators line up between rows. reStructuredText is particularly pedantic about such things.
With MoinMoin, whose native table syntax does not insist on some kind of decorative compatibility with a plain text representation, the single line per row restriction still applies. Obviously, this constrains the kind of content you can put in tables rather significantly. Since I think MediaWiki's tables are useful, I ended up making an extension for MoinMoin that provided broadly the same level of flexibility, and this proved useful when targeting MoinMoin in format conversion exercises (in my case, I took on a task converting from Atlassian Confluence).
I imagine that this exercise gave you some idea of how often these format mismatches arose and whether content fidelity would suffer if continuing to target Markdown. Previously, it was noted that HTML could be used as the target format, which would potentially be less convenient in some ways if the content were to remain editable in a similar fashion to that originally provided, but it might be the best compromise in the end. I don't know what else GitHub supports, if anything. In a work context, I had to use Bitbucket where reStructuredText was said to be supported alongside Markdown but where the rendering was deficient and Atlassian made it clear that they didn't care, so I regard feature promises with skepticism.
Re: retrosoftware.co.uk archive
Hi all, retrosoftware.co.uk should be back up and running now!
Re: retrosoftware.co.uk archive
Nice, great job ......
Greetings
Kees
Greetings
Kees
- TobyLobster
- Posts: 622
- Joined: Sat Aug 31, 2019 7:58 am
- Contact:
Re: retrosoftware.co.uk archive
Nice work getting this back online, it's a useful resource. I've already added a couple of the multiply routines therein to my comparison page of multiply routines: https://github.com/tobyLobster/multiply_test/, now comparing 116 different multiply routines.