@shedd - Let's say you work at a large company that's stuck one version back on jQuery (only 1.11/2.1 are supported) or Python (only 2.7.6 of the 2.x versions are supported) or Drupal (only 7/8 are supported). It looks to me like you're out of luck.
Perhaps this makes sense to be a feature of the paid version, since smaller projects are unlikely to be stuck on older versions. (And I actually like the subtle push to upgrade)
This is one of my secrets to getting things done. I like to shift into "offline mode" fairly often. This alleviates most of the worries of not having Stack Overflow for basic code questions.
Interesting! Thanks for sharing @kapeli! I have a couple thoughts:
Re Implementation: You seem to have versions implemented unofficially now, via separate documentation packs. (eg It looks like Drupal 7 and 8 aren't "versions" of a master "Drupal" documentation, they're just separate top-level packs). This seems like a fine model in terms of data model and UX. Perhaps you could extend this to all version support, with older versions hidden by default. (and/or only available in the pay version).
Re Sourcing: You only need to scrape older versions once right? Perhaps the Internet Archive would solve this problem. eg, this appears to be the jQuery 1.9.1 documentation https://web.archive.org/web/2013...
I see what you mean about an incomplete versioning system being potentially worse than no versions. It seems like you already have that though with Drupal, Zend, Python, etc, no?
So you scrape -> CDN yourself -> download via client? (With exceptions for iOS and Mac it looks like?)
More broadly than archive.org: What are the challenges involved in maintaining a product with a large web scraping component?
Report
I use Dash every day. It's awesome, thanks @kapeli!
Replies
CryptoPoops
Perplexity
CryptoPoops
CryptoPoops
Coursefusion