Update migrated blog slugs

[noatamir]

It would be useful to go through all blog slugs and update them, so they are all:

• Lower caps
• Separated by hyphens rather than underscores
• Don't have overly generic names

Some examples are listed in @gabalafou's comment here.

[gabalafou]

On the subject, while I'm thinking about it, we might want to add a redirect rule that captures all of the old links and forwards them to the new link, like so: /blog/year/month/slug -> /blog/slug. Or maybe rethink the way we've been doing blog slugs on the new site.

[noatamir]

Indeed! I was just asking @MarsBarLee about it on another PR because I think she mentioned that the CI is adding them or “fixing” redirects and I saw no evidence of that.
Can we prioritize it since there are a lot of blog2blog links that need manual updating otherwise?

[bskinn]

the CI is adding them or “fixing” redirects and I saw no evidence of that.

Agreed, CI is not doing anything of the sort that I can tell. I've been adding all my redirects manually to next.config.js.

Regardless of what decision is made on the Labs blog slug paradigm, I'm thinking that it would be good to implement something cleaner for "simple" redirects like these for the blogs (esp. for Consulting blog), where for now each redirect is an explicit entry in the redirects entity in next.config.js. It would be more clear, simpler, and more DRY to have a separate file... simple-redirects.js or whatever ... that had all the redirect definitions in it, and then code in next.config.js that pulled that new JS file in and used its contents to autoinject the needed suite of redirects into redirects.

If folks like this, I can create an issue for it.

[gabalafou]

Just to make sure we're on the same page, you're proposing an explicit redirect table like the following, right?

Old URL	Redirects to
https://labs.quansight.org/blog/2019/08/spyder-40-beta4-kite-integration-is-here	https://labs.quansight.org/blog/spyder-40-beta4-kite-integration-is-here
https://labs.quansight.org/blog/2021/09/not-a-checklist	https://labs.quansight.org/blog/not-a-checklist
and so on...	and so forth...

That seems like the simplest way to do it (and possibly least error-prone, despite the volume of entries). I'm +1 on that.

[bskinn]

Yes, exactly that semantics @gabalafou. Though probably it would exist in code as a mapping?