PLoS One and a failed comment system

Why is the comment system of PLoS One a failure?

Simples, comments are essentially invisible.

This is the single biggest failure of the comment system. It is the main reason that so few papers have comments, and why so few people comment on papers in PLoS One.

So what does 'invisible' mean here? Well, the only way to find a comment on a paper, is to visit that paper and look at the comments tab... Well... what more do you want? Actually, you want a lot more. You want 2.0!

First, from the perspective of a commenter, I want to see a list of my comments. I want some place where I can see my contributions listed. I want something I can link to, and somewhere I can link people. I want my comments to be 'visible' in this trivial sense at least. I'm actually stunned that this feature hasn't been implemented yet.

Second, from the perspective of a reader, I want to see which papers are being commented on. Although I now see a small 'Community Discussions' section 'hidden' at the bottom of the homepage, there is virtually no way to see, or heaven forbid, track, commenter 'activity' on the site.

• I can't subscribe to replies to my comments (come on!)
• I can't subscribe to comments for a paper
• I can't subscribe to comments for a journal
• I can't subscribe to comments for a search term

By subscribe, I mean by either via RSS or by email notification (neither is possible).

People want credit for comments. People want people know that they've taken time to contribute. I want each comment I make to be tweeted, and I want to know when someone has replied to my comment. I want to see which papers are being talked about this month. I want to see where the controversy is and which papers are creating a buzz. Which authors get the most comments and in which fields? Which comments get read the most?

These things are basic to making a community commenting system work.

But I think there should be more...

I want rich comments that have all the qualities of a publication, including figures and references. I want comment systems that support computational argumentation so that debate can be managed. I want commenter credits when people rate my comments. I want to see who has made the most comments, and who has the best rated comments.

I have really no idea how PLoS One put so much stock in their lousy comment system, and how comes they haven't improved it.

If you know better, I'd be glad to hear it, because currently, it's embarrassing.

Downloading information from 23andMe using wget

I want a backup of 23andMe for when my account runs out, so I decided to use wget to download the various analysis available. Note, you can download your SNP calls from 23andMe, but the value the site provides is the associated interpretation.

First, lets just grab the extensive information that they already make publicly available (no account necessary):

wget -r -l 2 --include health/ https://www.23andme.com/health/all/

• -r - makes wget recursively download pages linked to the given URL.
• -l 2 - keeps recursion to a maximum of 2 levels, which is all we need I think.
• --include health/ - only downloads pages with 'health/' in the URL, which is all we want in this case.

Recursive wget can be 'dangerous', because, essentially, it starts downloading the whole internet, following all links it finds. Here we just grab pages that are two links 'deep', staring from the given URL, and only links under the 'health' directory. In this way we never leave the 23andMe domain, and we only grab what we're interested in, ignoring all the other junk that makes up a web page.

The result won't be navigable (see -m for a more 'robust' solution), but I don't really care, it's the information I'm after, not the layout.

Note that the display of the resulting HTML in your browser will look pretty because it uses absolute links to css, js, and images. This won't last when 23andMe take the information down. i.e. if you really care about the result being pretty, see wget -m.


Now to download the data in my account...