Underscore and Lodash Merge Thread
github.com
github.com
From the author: "Folks. Please STOP +1 and :+1: -ing GitHub tickets. We get it, but it makes tickets hard to have conversations on. Contribute with something substantive instead."
This is a very common problem for github issues posted on HN/any social media. Would be great if github had a voting mechanism.
Because subscribing does not show interest/support: the repository owner does not see how many people have subscribed to the issue. Furthermore, supporting the resolution of an issue does not mean you want to be spammed with updates to said issue.
> Google's issue trackers have the same issue: they support starring but people insist on doing "me too" posts.
Google's issue tracker has the same issue to a lower extent: while everybody can see the number of stars on an issue (which is good), starring an issue requires following it and being spammed by it (which is bad).
Then again, people are by and large idiots, even on e.g. youtrack (which has non-subscripting voting) +1 comments abound: http://blog.jetbrains.com/youtrack/2013/01/making-1-mean-wha...
Only way I see to mitigate that is assuming the tracker has orthogonal subscription and voting ruthless moderation and removal of worthless "+1" comments.
Commenting on the issue automatically subscribes you to it. You have to manually unsubscribe if you don’t want to be spammed.
The +1s are so irritating because you have to read through a huge conversation and you also keep getting email notifications because you want to keep in touch with a thread. It actually makes it harder to do what the github topic is about with the +1s.
They should have locked the thread or use a wiki page for this. A wiki page is less likely to be edited to add this kind of comment or at least it would be easier to clean up with a bot.
It's surprising that Github doesnt provide "approval" as a native feature
... :)
It's annoying in some cases now because 1) it's anyone, not just the active contributors or even users, 2) people don't use it as a vote but as a general expression of support, 3) -1 isn't really used anymore so you can't scan them and see what the sentiment is, and 4) in this case, it's not even clear what they are lending support to- the merge, or that it is time to discuss a merge seriously.
So yeah, in this case it's noise but in some contexts +1/-1 is a quick and simple way to govern code and it has historical roots. Just ranting as I think they and some in this thread are a little too quick to judge. +1/-1 from the right people could be considered very relevant information to a discussion (esp. within a focused community).
Never knew that before. Does Lodash have more features than Underscore? I never got too heavy into using either but I always thought of them as essentially being interchangeable libraries.
It's not a macro framework like jQuery... the kitchen sink isn't in there... though, I do think that probably the string templating could/should become a sub-module... the rest belong together more.
A few months old but should still be reasonably accurate
Yes Lodash has some interesting 'better than native' optimizations that utilize JIT compilation
I very much appreciate being able to use Ramda for composition of methods, which can mean much more concise code[1].
[1] https://github.com/tracker1/azure-config-util/blob/master/sr...
At least mind share wise I've heard of lodash/underscore much more often than ramda. Only recently has it come up with some level of frequency.
Is there anything else about it to that makes it really "eclipse" them or is it more like "ramda is also nice, especially because it has Currying"?
To me it's more like a nice-to-have -- a helper library of functions you could write yourself (but probably shouldn't bother to).
Other than perf, modularity, etc, I like lodash's doc better. The doc is written as inline jsdoc in lodash.js, with each parameters documented and behaviors explained through examples.
But here is @jashkenas's take:
> Documentation style. Ideally I'd like the Underscore docs to be written even more conversationally than they are at the moment.
Personally I also trust @jdalton more. Other than technical skills he's very hard-working and responsive. He seems to be following a zero-issue-and-PR policy and get all issues/PRs sorted out in 2-3 days.
This reduces code noise and amount of typing if you are using these functions frequently
I'm not sure how that is silly or arrogant
The problems with using single-character names for libraries are so self-evident, I'm not quite sure how to respond to someone who doesn't see it.
* list and dict were lacking vital methods like list.remove() and dict.__len__()
* dicts weren't really a real thing (real ones coming in ES6 https://developer.mozilla.org/en-US/docs/Web/JavaScript/Refe... ), but it was easy to add random properties to arbitrary objects while accessing a missing one simply returned None, and so these were used instead
* Which leads to for x in dct being subtly broken in certain use cases unless you understand the above and how it interacts with prototypes
* So loads of utility functions were added to deal with the above facts, plus the kind of useful tools you would find in Python's functools and itertools and a large number of other utility functions while they're at it.
(EDIT: One of these days I will remember HN doesn't use Markdown, but not today)
Apparently underscore is a community-driven "stdlib" that implements more functionality and saner container classes, and has more freedom to evolve since it's not part of the core JS language.
And the justification for using a single-character name is that it's basically a de facto part of the stdlib?