It is funny how they do things, they are really established and they can do things totally their way, two years without that feature could be inmoral for other companies.
These guys produce an entire awesome startup's worth of development every couple months. Definitely need to be paying more attention to the how of what they do. Luckily Tom's quite the excellent speaker and writer. When he talks, I listen.
I agree. As soon as I can afford to, I am doing the same. This is definitely a service, that I don't necessarily use a lot now, but if they went away I think the industry would suffer for it.
Everything they do inspires me to continue pressing on.
I don't totally agree with the "almost nothing". I was already using them for all my project and it was just enough for me. The second you have labels, you can mostly do any arbitrary complex structures.. But yeah, for medium to big project, that would not have been enough. Now with the new issues system, it might work.
Scott Chacon is also excellent in this area. He was a Codemash keynote presenter in January of this year. He talked specifically on how GitHub is organized, how they decide what to build next and how they get those things done.
Anyone care to give their thoughts on GitHub Issues 2.0 vs Pivotal Tracker?
(sorry, I know question-in-a-question is a little poor form, but I only really care for the opinions of the HN community on this rather than asking more widely)
I haven't used Pivotal, although I understand Pivotal is more of an agile project management tool. It measures burn down rates, velocity etc... even if it does do issue tracking as part of that, it's probably not reasonable to compare.
A like-for-like issue tracker would be Jira or Lighthouse
Indeed Pivotal Tracker and Github Issues focuses on two different things - Pivotal is about planning, Github Issues is more like a registry. They may actually complement each other although I haven't checked if Pivotal provides integration with Github Issues already (it certainly does with Jira and Lighthouse).
I think this is a pretty good point to bring up the comparison. I can definitely see github ousting many software dev tracking tools in the years to come. Very interesting...
GitHub should consider adding planning and velocity tracking tools if they really want to unseat Pivotal or other "Feature Planning" tools. I'd switch in a heartbeat just as a means of consolidating my tools.
If you want something for real planning of future development it doesn't look like issues 2.0 really fills that need.
That's unfortunate because issues 2.0 is nice for what it does but pivotal also does issue tracking in a way that is really well integrated into my overall work planning. Using both probably doesn't make much sense.
So although these features look nice by themselves, I think I'll continue to not use github's issue tracking in favor of tracker.
I chuckled when I read the word "triforce" near the beginning of the article. I assume they're either referring to the Legend of Zelda or an offensively-named 4chan meme.
This is great. Milestones and assignees are the only things I've found myself missing in Issues besides file attachments.
Now I just wish there were a way to mass edit issues without checking each one individually. It's going to be a pain to migrate every issue in our 1.0 tag to Milestone 1.0.
I don't run pentadactyl but it looks like it supports vim-style macros. I assume it would be something like qAjx to create the macro and 25@A to use it if j and x pass through to the page properly.
I had been using an @name label for assignment, it's nice to see this supported. But filw attachmentes are still missing right?
This is a really important feature that makes issues much less useful than it could be :(
Same situation here. I’d love to push our company to Git+GitHub and the new issue tracker is another small advantage in comparison to our current setup, but it’s a no-go until we can attach screenshots and logs to the issues. Let’s hope that file attachments simply did not make it into 2.0 and are scheduled for one of the first point releases.
While I really, really like the new Issues 2.0, I think the biggest news I got out of the article is pointers to PJAX. I may be behind the curve but I've never heard of it and its so cool I think I'll use it in my next project. :)
hmm, am I the only who wishes that issues where simply files in a directory inside the repository?
It seems crazy that when you fork an old version, you have no idea which bugs have been fixed and if you later merge again you have to manually track down which bugs have been fixed.
Is there some up to date comparison of github / bitbucket now? It seems like they fixed the issues view already and the last big difference is the number of private accounts allowed without paying. Is there any other outstanding problem with any of them?
Unfortunately, Kiln was never an option for us. We are a git shop, and while I do have a lot of respect for hg, we just do not want to use it. But VCS integration in FogBugz was never the main problem.
We currently use FogBugz with GitHub (private repos) and it works very well. I wrote a bridge that connects both APIs together, so we get fairly good integration. Information about commits and branches gets attached to FogBugz cases and we get a smooth workflow out of it.
The problem with FogBugz is that it is slow and has a number of minor annoyances, which grow to be a pain in the long term. Such as Fog Creek refusing to make the software automatically mark "#1527" as a linkable case. You have to write "case #1527". Quite annoying, especially if you don't run your FogBugz installation in English!
There are other things, such as obsolete Windows-based character encodings they insist on using in E-mails sent from FogBugz. Also, hosted FogBugz is slow. We spend a lot of time waiting.
Oh, and the default "VCS integration" features in FogBugz are from the CVS/RCS era and they still haven't been modified, even though they sell Kiln. The "Checkins" field in every case is file-based, so there is no way you can attach modern DVCS changesets to it.
I reported all of the above problems more than a year ago, and nothing has changed.
All in all, while we do use FogBugz (and pay quite a bit of money every month for the pleasure), we are looking for alternatives.
The milestones are a welcome addition, but the removal of priorities and voting is a real pain. I had a bunch of issues painstakingly sorted into priority order and they're now essentially shuffled.
The UI is still a little bizarre as well - no bulk select? No way to go to the previous/next issue from the actual issue? Colours of tags not displayed in the issues list view? I can sort by number of comments but not order them myself? I guess some of these might be considered bugs which will be fixed with time but others are really strange choices which don't quite seem to fit with the rest of the GitHub UI.
Check one issue, hold shift, click one further down.
Ah, I see - I hadn't tried that, thanks. The old select all/none would be great to add back.
I see the colours on the list of labels to the left of the issue list, but not in the labels themselves on the actual issues (see https://github.com/cmf/schemely/issues). Additionally, when I select the label as you suggest the label always seems to be shown in the dark red that appears in the issue list.
While I'm griping, the migration to 2.0 also appears to have lost the ordering on my tags, and I don't seem to have any way to order them - this is particularly noticeable since I have 3 severities of bug and 3 classes of feature request.
I noticed a couple of other bugs as well - is there an issue list for GitHub itself we can file bugs in?
I love the attention to detail such as their label foreground/background colors when selected (it seems that some have white foreground, some have "letterpress" effect, depending on the contrast).
55 comments
[ 0.21 ms ] story [ 131 ms ] threadAnyway, welcome new issue tracker!
I just got a GitHub paid account, not because I need private repos but because the company is so darn awesome.
Everything they do inspires me to continue pressing on.
HN discussion: http://news.ycombinator.com/item?id=2110227
(sorry, I know question-in-a-question is a little poor form, but I only really care for the opinions of the HN community on this rather than asking more widely)
A like-for-like issue tracker would be Jira or Lighthouse
That's unfortunate because issues 2.0 is nice for what it does but pivotal also does issue tracking in a way that is really well integrated into my overall work planning. Using both probably doesn't make much sense.
So although these features look nice by themselves, I think I'll continue to not use github's issue tracking in favor of tracker.
Now I just wish there were a way to mass edit issues without checking each one individually. It's going to be a pain to migrate every issue in our 1.0 tag to Milestone 1.0.
The fastest way I can find to do it is through the keyboard shortcuts.
http://dactyl.sourceforge.net/pentadactyl/index
What does that have to do with the keyboard shortcuts on github's home page?
It seems crazy that when you fork an old version, you have no idea which bugs have been fixed and if you later merge again you have to manually track down which bugs have been fixed.
https://github.com/bartdag/gitli
I'm glad, because I think Fog Creek has stagnated and FogBugz isn't developed much anymore — competition is good for everyone :-)
We currently use FogBugz with GitHub (private repos) and it works very well. I wrote a bridge that connects both APIs together, so we get fairly good integration. Information about commits and branches gets attached to FogBugz cases and we get a smooth workflow out of it.
The problem with FogBugz is that it is slow and has a number of minor annoyances, which grow to be a pain in the long term. Such as Fog Creek refusing to make the software automatically mark "#1527" as a linkable case. You have to write "case #1527". Quite annoying, especially if you don't run your FogBugz installation in English!
There are other things, such as obsolete Windows-based character encodings they insist on using in E-mails sent from FogBugz. Also, hosted FogBugz is slow. We spend a lot of time waiting.
Oh, and the default "VCS integration" features in FogBugz are from the CVS/RCS era and they still haven't been modified, even though they sell Kiln. The "Checkins" field in every case is file-based, so there is no way you can attach modern DVCS changesets to it.
I reported all of the above problems more than a year ago, and nothing has changed.
All in all, while we do use FogBugz (and pay quite a bit of money every month for the pleasure), we are looking for alternatives.
The UI is still a little bizarre as well - no bulk select? No way to go to the previous/next issue from the actual issue? Colours of tags not displayed in the issues list view? I can sort by number of comments but not order them myself? I guess some of these might be considered bugs which will be fixed with time but others are really strange choices which don't quite seem to fit with the rest of the GitHub UI.
Colors of tags are definitely displayed in the issues list view - right to the left on the label itself, or full width in the case it's selected.
Ah, I see - I hadn't tried that, thanks. The old select all/none would be great to add back.
I see the colours on the list of labels to the left of the issue list, but not in the labels themselves on the actual issues (see https://github.com/cmf/schemely/issues). Additionally, when I select the label as you suggest the label always seems to be shown in the dark red that appears in the issue list.
While I'm griping, the migration to 2.0 also appears to have lost the ordering on my tags, and I don't seem to have any way to order them - this is particularly noticeable since I have 3 severities of bug and 3 classes of feature request.
I noticed a couple of other bugs as well - is there an issue list for GitHub itself we can file bugs in?