Giters

cscheid / rgithub

R bindings for the github API

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

confusing that GitHub repo `rgithub` = R package `github`?

jennybc opened this issue · comments

commented

Perhaps I am the only one to get tripped by this?

But since the repo is named rgithub and the package one will immediately want to load is github, maybe you should mention this in the README.md.

My install and explore sequence started like so:

> library(devtools)
> install_github("cscheid/rgithub")
Installing github repo cscheid/rgithub@master from cscheid
<blah blah blah>
> library(rgithub)
Error in library(rgithub) : there is no package called ‘rgithub’
<much puzzlement … listing my libraries … AHA!>
> library(github)

Looking forward to using the package!

commented

Yes, that’s a very good point. The naming of this thing is awful and my fault.

the package is “github” because it felt redundant to type “rgithub” in R, but naming the repository github felt wrong because the url would be extra weird: http://github.com/cscheid/github.

I have come to believe that I picked the maximally bad tradeoff.

In any case, the package needs renaming before I send it to cran (“no trademarks! we don’t care if it’s ok with the parent company!”), so I’m taking suggestions :)

Sorry for the confusion!=

commented

How about rhub? hubber? (Better not check the urban dictionary....) rgh?

I don't think any of these are particularly good, just wanted to throw in something to have a baseline.

Btw. "no trademarks" is interesting. How about packages like RMySQL, Rmongodb, rentrez, etc? Surely, adding the letter does not eliminate trademark protection.... MikeRoweSoft could talk about this.

commented

I agree with "no trademarks" being silly. You’re not arguing with me here, you’re arguing with the Powers That Be :)=

commented

I know, sorry, I was just thinking aloud.....

commented

Depending on your tolerance for puns, you could take Github's "octocat", stick an R in front of it to get "roctocat", and try to find some play on words. If you pronounce that thing out loud, the Clash song comes to mind. (Octocat is also a trademark from Github so we can't use it directly)

Is there a way to say "eight-legged feline" with fewer than 17 letters?

commented

roctopuss ? Or you can just abbreviate it as russ.... But roctocat is actually good, imo.

commented

(can’t use roctocat...)

but a multitude of cats and R made me think of https://twitter.com/hashtag/rcatladies

maybe we should conscript the help of @hilaryparker or @karthik=

commented

Jokes aside, it would be good to have a name that reflected the functionality of the package itself. So I 👍 @gaborcsardi's suggestion for either rhub. Even if roctocat is ok with CRAN it would be worth checking with GitHub too. I know they do now allow naming apps with their trademarks.

commented

I like roctopus too.

commented

I think the overall discoverability of API wrapper packages would be greatly enhanced if there were an actual convention. Dry and boring, I know, but useful. We could still get cute by spelling wrap as rap or rrap. 😄 So rap_github, rap_<whatever>, ... Then they would all show up together naturally in lists of packages, etc.

BTW I actually got this package to do my bidding early this morning, so that was terribly exciting.

commented

@jennybc, that is a very good idea. The bikeshedder in me wants to solve this by creating a metapackage that generates wrappers from something like a json description, but I really don't want to reinvent SOAP.

Do you have a list of packages you think should all be named consistently?

commented

A lot of the stuff I see here:

https://github.com/ropensci/webservices#data-sources-on-the-web-accessible-via-r

I knew that train had already left the station ….

commented

Picking this up.... @karthik, if I choose roctopus I'm going to get yelled at: http://www.rforge.net/ROctopus/index.html :)

commented

Just for completeness: roctocat is a non-starter not because of GitHub. GitHub allows libraries to use their trademarks, but CRAN's policy explicitly forbids it. It's precisely what bit me with github. I'm not inclined to go argue my case again to the CRAN police..

commented

I don't know if CRAN complains about stuff on rforge! But yeah arguing with CRAN police is pointless.

http://cran.r-project.org/web/packages/ROctopus/index.html

:)

commented

Oh, in the case of ROctopus, it's @s-u who'll yell at me ;)

commented

@cscheid Is there any place exactly in, e.g. CRAN policies, that makes you think rgithub or githubr or whatever wouldn't fly? I care for the sake of this package and also my wrapper around Google Sheets. I mean, we have googleVis, after all!

commented

Um, there's a certain someone that controls the acceptance over at CRAN. And he gets the final say. And my submission got rejected. And I argued that there were no fewer than 28 other libraries using GitHub trademarks officially endorsed by GitHub. To no avail.

commented

Dear @mages and @JCheng,

Can you share any wisdom about getting googleVis onto CRAN with that name?

Thanks.

commented

There is also leaflet going up soon for e.g.

commented

@jennybc Here is the vague wording on the policies wrt trademarks.
http://marker.to/ntMRjh

commented

I have much respect for GitHub. And Google.

commented

Well, perhaps I got away with it because 'to google' is actually a verb in the Oxford Dictionary and googleVis is just a play on that.

commented

Ok, any downvotes for hubr?

commented

grithub, gitr, gitub, hubub

just checked: https://github.com/melix/grithub

if you want to get cute, even hubby can work

commented

gitr since there is git2r, another possibility is github2r ... but then you're back where you started w/ GitHub in the name 😕

commented

+1 for hubr. note that rhub would also be a conflict (and not just with the delicious chard).

commented

@jennybc: turns out @soodoku did suggest grithub. I like that one the best right now. I'll get started on this when I get back to Tucson.