Hacker News Reader

Comment by CydeWeys | Hacker News Reader

I do see why they ask for that (although it sounds like their systems could use some improvement).

Imagine you trying to get someone to review a 1,000 line patch. If you can just get them to log in to GitHub, they can use all of the pull request commenting tools, and it'll be easy. Otherwise, you have to mail them a big .patch file, and then somehow manually manage comments across all of that. The assistive tools are more convenient.

View on HN

And add that you have dozens of such patches that you need people to review, and their reviews have to be distributed to a committee of senior developers once they're received, and so on and so on...

A journal may have dozens of submissions at various stages of the peer review process. Each requires two or three peer reviewers, who are busy academics who will often fail to meet the deadline. The editor has to track all of these, follow up with lagging reviewers, and then put together all the reviews for a paper to make the final decision.

Doing all that by email would be a mess. Reviews would get lost, the editor would lose track of which reviewers were sent which papers, and so on.

So the online systems make sense. But they all suck. I've reviewed for several journals which use the ScholarOne system, for example, and despite all being hosted centrally, they inexplicably require separate logins. The website is a clunky piece of crap. If they just had an automated system email me a PDF and ask me to email back a review, and had their script grab reviews from the inbox and format them for the editor to review, it'd be a lot easier. Or if they emailed me a personal link that let me view the paper and leave a review, no login required. But alas, they must overcomplicate it...

Linux kernel patch review is over email like this. It seems to work. I subscribe to a few lists and it makes it easy for anyone to chip in.

It's a bit burdensome on the patch submitter, to collect the feedback into one place, but not that much. At least the burden is on people who want to get the patch in.

Github accepts Google account sign ins. I personally no longer register for services that do not accept my google account as a form of authentication. I wish not to leave even more online accounts and passwords scattered over the web. With google auth, they do not get a users passwords. Without google auth chances are most people will fall for password-reuse.

Last time I saw a Google account log-in, the remote service wanted full access to my contact list. That is, it was demanded that I betray third-party confidences to friends in order to gain an authentication to the service.

Fuck no.

Here: https://ello.co/dredmorbius/post/vv0bq6oia_06z_yjnmjwzw

I wish we still had Persona.

But there aren't twenty GitHubs. There are three or four, and their are hosted versions of those that are slightly different.

If publishers all used the same software package for content management, we might be somewhere. But every company with an IT department seems to build their own Content Management System at some point. I joke with coworkers that we would solve the developer shortage in north america if we figured out how to stop everyone from writing their own CMS.

So then give them a one-time login ticket, and let them upgrade to an account?

Not saying you could do that with GitHub, but there's design space for it.

Reply on news.ycombinator.com