Rik

@greggraf gonna check this out ASAP, but just to say thank you so much for persevering with this, it's a huge contribution.

@mafintosh the above error is being reported by a small number of clients. They are using the hyperdrive:`07e82e6ea1ab7e7c3d4872f035b2320e8aefe26ba35033b983f4b2e033d07d44`, which definitely has en entry for `/articles/2/0/6/7/4/elife-20674-v1.xml`. Any suggestion as to why...

Note, this is only happening on linux clients

Curiously, this paper and the associated data are missing from the original dataset. I am not quite sure how they came to be indexed in the dat, but subsequently removed...

@billbrod OK no that sounds like a different bug. Sorry for the poor experience. Investigating this now too!

(old note in the Keyboard shortcuts project converted to an issue in light of #130)

Mousetrap has now been added by @CAYdenberg in #130. There's still some discussion to be had around how to generalise registration of shortcuts, but for the time being keyboard shortcuts...

Looks like #92 was not resolved - though I haven't been able to reproduce this. Sorry for the trouble @gmaciocci - we'll figure it out! If anyone else experience this...

I'm starting to think about this in detail. Here's what I've got... - We should choose between **packages**, **extensions** and **plugins**. These names are all used in the wild in...

Looking at the pre-post middleware hook pattern, for which there are a bunch of fairly simple packages: - https://github.com/bnoguchi/hooks-js/ - https://github.com/indexzero/broadway - https://github.com/bmeck/understudy