larsks / blog.oddbit.com

3 stars 0 forks source link

post/2019-05-07-new-comment-system/ #2

Open utterances-bot opened 5 years ago

utterances-bot commented 5 years ago

New comment system · The Odd Bit

As long as I’m switching site generators, it seems like a good idea to refresh the comment system as well. I’ve been using Disqus for a while, since when I started it was one of the only games in town. There are now alternatives of different sorts, and one in particular caught my eye: Utterances uses GitHub issues for storing comments, which seems like a fantastic idea. That means that comments will finally be stored in the same place as the blog content, which I think is a happy state of affairs.

https://blog.oddbit.com/post/2019-05-07-new-comment-system/

rickysarraf commented 5 years ago

That is interesting. Do the comments become part of the content ? I mean, if tomorrow you move on to a different comment system or a different site generator, do your comments get preserved ?

larsks commented 5 years ago

The comments are simply comments on a github issue. If I were to move to a different site generator I could continue to use Utterances. Exporting comments for use in another comment system would be trivial (the GitHub API is quite powerful), but the ability to import them into another comment system would really depend on the import capabilities of the target.

rickysarraf commented 5 years ago

Great. But what I've sorely been missing is a comment system which could simply attach/append the comment to the content itself. That way the data remains in the hands on the owner.

Over the years, I've suffered comments loss. This data is valuable but as it turns most cms/blog give comments a lower priority. And migration tools further lower that.

And this hurts especially when you decide to move your site to something different, after your current tool is outdated/unmaintained. In such circumstances, desperation to move to a more current and maintained tool is desired. As in doing that, many data is lost. Comments, Taxonomy, Statistics

larsks commented 5 years ago

Huh. I don't want the comments appended to the content :). I'm happy with having the comments stored in GitHub: I feel that the data does remain "in the hands of the owner" this way, certainly compared to other solutions. It's easy to maintain a backup of the comments outside of GitHub, for example.

The problem with comment migration is that in most cases, comments are authenticated: so, e.g., I could export older comments from Disqus and associate them with the appropriate blog posts, but they would all appear to be from me rather than the original commenter. I don't think there's a way to solve this problem without actually running your own comment engine, which is exactly what I'm trying to avoid.

VincentTam commented 5 years ago

The comments are simply comments on a github issue.

@larsks Hi, this is the maintainer of @staticmanlab, a public GitLab instance of Staicman. IMHO, it's better, in terms of SEO, to treat comments as part of the site's content instead of problems of a software package.

Staticman makes use of GitHub/GitLab Pull/Merge Requests instead of issues. Under Staticman's model, static comments are YML/JSON files stored in the remote GitHub/GitLab repo (usually under data/comments, configurable through the path parameter in root-level staticman.yml), and through a static blog generator (Jekyll/Hugo/etc), the stored data are rendered as part of the content. This gives a total ownership of a static site's comments.

Moreover, Utterance and Gitment only support GitHub, whereas Staticman supports both GitHub and GitLab. Utterance and Gitment require a GitHub account to comment---this can hinder many non programmers from leaving comments to your posts. Staticman allows both authenticated and unauthenticated comments, so it's possible that commentors leaves a comment without any login.

larsks commented 5 years ago

Vincent, thanks for the pointer. Looks like an interesting project. I think I have mixed feelings about making comments part of the actual repository content, but I can see why some people would prefer that model.