Reorganize the Archive

[nikhiljha]

Reorganized the summary page based on Wikipedia's list of algorithms.

Also working on reorganizing the content itself, so WIP.

[leios]

I agree on re-organizing the summary page and such; however, once we decide on which scheme we are working with, the urls will change and I will need to update all my links on youtube (and other places) to match the new archive. In addition, we'll lose our search presence because of it.

I tried to get this conversation going in #57, because it's incredibly important... but there wasn't much discussion then. Ultimately, whatever scheme we go with, we need to stick to.

The idea of the Algorithm Archive was to provide a cohesive curriculum for all areas of computational science. We can still do this, regardless of the directory scheme by providing "guides" for (for example) computational physics, which provides links for all the chapters in the area and secondary commentary.

[nikhiljha]

the urls will change

Luckily, I don't think so. All I moved around was the SUMMARY.md which as far as I know doesn't affect the URLs. Not 100% sure though, someone should test this.

[leios]

Oh, you are right. As long as we don't change the directory scheme, we are fine; however, I would like to have an introductory blurb to each area. I suppose we could have a different SUMMARY.md to the directory scheme. I feel like if we change the directory scheme, all the PR's will have merge conflicts.

[nikhiljha]

Oh yeah, didn't even think of the merge conflicts. Resolving those merge conflicts should be super easy (in GitKraken at least, which I use exclusively for fixing merge conflicts), just move the file to where it needs to be.

And as far as introductions go, we can just make a file called "Introduction" under every heading. That also gets around the limitation of only being able to have single level subheadings.

[Butt4cak3]

You added a bunch of commits that are unrelated to the PR. Please try to work with branches and keep PRs focused on one topic. Everything else makes them difficult to review and merge.

[jiegillet]

I'm not in favor of this PR. It is quite rushed, and I would leave jobs like reorganizing the AAA to @leios. Start with contributing by submitting or reviewing code and getting to know other members, not by tearing the house down.

[leios]

To be fair, I do think the chapters need re-ordering and I think something like this makes sense. I just don't know the best scheme to go with for now.

[Butt4cak3]

I must admit I'm with @jiegillet on this one. In addition to that, we should first discuss the matter, come up with a solution and then start to work on it, not the other way around.

[nikhiljha]

Of course, you can still discuss the matter first. This is just an option that you can choose to merge or not.

Also as far as unrelated commits, I removed a few paragraphs/words that were old and no longer correct. I think that fits with the theme of reorganization/cleanup.

[june128]

Even if we decide to make this change, I strongly suggest to update the directory structure as well and add redirects for the old links. But before we put in all this work, let's discuss this topic.

[nikhiljha]

The directory structure seems fine. Everything is a root directory, which will help whatever other "custom books" that are planned to do later.

[nikhiljha]

And I think we are having a discussion here, what other options are there?

[leios]

So we have some pseudo organization (for example sorting and searching and physics solvers), which might conflict with other organizational schemes. No matter what organizational scheme we go with (maybe even allowing all algorithms to be in the main directory, while only touching the SUMMARY.md file for re-organization), we'll have to modify some directories.

If all the algorithms are just sitting in the main directory, then the urls will never change no matter how we change the SUMMARY.md file. I don't think that's a bad solution. It does mean that we'll need a directory filled with chapter introductions (that are not algorithms) in their own directory which can be added to the summary as need be.

[june128]

I actually like the idea of a very flat directory structure. It would make the future-proof regarding potential reorganizations. A directory structure complying to the website structure is obviously nice as well.

[nikhiljha]

Yup, that's what we're working with right now. It's a flat directory structure and a modified summary.md.

Our current pseudo-organization also fits with the Wikipedia organization, so that works.

Are there any other ways we could organize things?

[leios]

I meant that we need to remove that organization too and create a separate introductions/ directory or something and have an markdown introduction file for each heading.

If we are going for a flat structure, we need to move all the algorithms to the base directory, in my opinion -- Especially because we are already muddling things up with the physics solvers directory. It's not clear what the difference is between a physics solver and diff. eq. solver

[june128]

Yeah, so a flat directory structure could consist of all algorithms having an own directory just below base. Then having all the introduction markdowns in an introductions directory seems good as well.
The only special case is the principles_of_code thingy tho. Maybe we make everything in there flat as well? Just having priciples_of_code with all mds in it.

[leios]

To be honest, the principles_of_code section might need a rework as well. The way I see it, we can have that as an exception for now

[june128]

I mean making it flat wouldn't be much of an issue.

[june128]

I would suggest making the reorganization after #170 tho, since I add a few res folders there.

I should start finishing that...

[june128]

Ok, this PR implements flat hierarchy and therefore paves the way for a reorganization: #242

[leios]

I think part of this PR was solved with the PR #242; however, the underlying issue of the SUMMARY.md being unruly still hasn't been addressed. This can now be fixed without touching the directory structure; however, since the directory structure was heavily revised, we might need to change the PR a bit.

[nikhiljha]

Yeah this PR is pointless now. That change can be addressed in a different PR.