GitHub's growth in 2013 has been fascinating to watch. They've been relentless with new functionality, and the overwhelming trend with that new functionality is a marked increase in the scope of their operations: stuff like generic data support, baked-in visualizations, and quality-of-life improvements points to the conclusion that GitHub is going to stop being the "best place to share code with friends, co-workers, classmates, and complete strangers" [1] and start being the best place to share your work and data with everyone.
Relentless with new functionality? Yes, no, maybe, I don't know.
I'd love to see things fixed up in their issue tracker, and you still can't apply a label to a pull request while looking at the pull request.
I love GitHub for many reasons, but I question whether the push towards 'managerless-culture' means some less interesting things don't get done sometimes.
I do note the ability to see web traffic prior to someone's implementation of a hack to do it on HackerNews, but on the other hand, I still miss the impact graph. How about being able to see number of clones and updates instead of just web traffic?
The ways to spotlight interesting new projects is great, but I think the social and discovery aspects could be upgraded too.
But #1, the bug tracker needs help. Code review tools seem nice, but are seemingly enterprise only. While I welcome and encourage that kind of business model, large open source projects could benefit the most from that kind of functionality.
So, yeah, GitHub is becoming the standard for code sharing. It's great... but there's a lot of opportunity not being capitalized on too relative to their organizational size.
From the outside, it looks like they are probably dealing more with internal plumbing than new functionality to me. Which is also understandable. It's just a shame that some features that are quite useful (say, impact graph) get cut or uninteresting to work on (issue tracking) don't get worked on.
The code review pales in comparison to Crucible. Vertical diffs only, can't ignore whitespace without losing the ability to comment, and there's no fast navigation through files.
Well, it seems to me like they have only been implementing closed mechanisms instead of standardizing stuff, using public resources and/or opening their code. I hope that I'm wrong and it's just my pessimism talking.
Github PLEASE work on some kind of 'Proof Of Activity'
98% of my work is private and all employers get to see is about 100 open-source commits when I have roughly 4,000 commits throughout several private organizations. I know I need more open source contributions but I should be able to prove to an employer that I am a heavy Github user and have more experience than 100 commits.
Sometimes I've been asked to send a screenshot of my total number of commits, and happily oblige to the odd request but a screenshot is not even worth it's weight in the physical world (aka nothing).
Using total number of github commits as any sort of qualifier for employment is as bad (maybe even worse) as asking for a wc -l LOC count on your last project.
To be clear, I don't have an issue with github contributions being seen as a plus for hiring... but trying to turn this into something concretely quantifiable via total commit count (especially on commits you can't even see) seems crazy to me.
Please let me know which employers are asking you for this information so I can avoid applying to them in the future!
They are not(I hope) judging quality of my work based on # of commits.
-They are assuming I have little to no experience in a service they use every day.
-They are assuming I don't contribute to Open Source in any way (I contribute to private repos that are at the core of several large OS projects but are private for business concerns).
-Github provides an easy metric to see what most of my work is. I can see that 45% of my work is done in Python, and maybe the rest in front-end tech. That would be useful information to an employer if I apply to an Obj-C position.
Actually the work I am most proud of is in my side-projects, which I would much rather employers go on and interact with rather than scour over the github repo. Posted above is not just something I want, but something I see much more value in for Github to become a more friendly intermediary between employer/employee.
> They are assuming I don't contribute to Open Source in any way (I contribute to private repos that are at the core of several large OS projects but are private for business concerns).
How are those open source? Or are they open source but not open to third party contributions? Or does development not take place on github? Just curious since if the code isn't public I'm curious how it's considered open source.
How would you handle this case if the work you're doing that's being stored on private GitHub repositories wasn't on GitHub at all (like the majority of tech work that employees do for their employer)?
The point is that it's not entirely useful for employers, which seems to be what you're concerned about. If 80% of people are committing to private corporate source control, and 20% are committing to private github accounts, it's not readily apparent why you would care to see some vague statistics about commit frequency for 20% of your candidates.
-They are assuming I don't contribute to Open Source in any way (I contribute to private repos that are at the core of several large OS projects but are private for business concerns).
If you're commiting to private repos like that, then you're not contributing to open source.
Sure, you're programming, and contributing to projects that others use, and yes, your github profile doesn't show that. But it correctly shows your level of open source contributions.
As someone who designs metrics for a living… Yeah, it would be… bad. But if they are fine with screen grabs, it's most likely a filtering system: real talent selection is at the interview level. For instance, I have no idea what to answer when they ask if I'm a developer (I ‘code’ in SQL and Python to set up BI metrics). If they ask for that kind of pre-requisite, I know what they mean and I shouldn’t apply.
Same for: what kind of editing software/UDI/Python or R package do you prefer? I'm not expecting them to judge how clueless I am by saying “I know scikit-learn is cool, but I prefer pyml” but rather, that I know what those are.
Having the bulk of my Github commits all going to private repositories, I can empathize with this.
For me, in 2013, I made close to 800 commits to private repos. When I'm logged in, I can see the activity in the "Your Contributions" section of my private profile. Last year, I also switched employers and because I don't have any public projects I contributed to, my profile looks barren. However, that's far from the truth.
It would be nice if Github would allow me a setting to display just the commit activity, but not link to the private repos. I wouldn't be surprised if this is far off.
For those that would say "you should be contributing to more OS projects then", just remember that there are some of us who can't code 14 hours a day (anymore). We have families and/or lives out side of coding. This doesn't make our effort any less valuable than those that do it for the public. We just have a different set of constraints and responsibilities.
My dissertation project is in a Github private repository, yet it looks like I abandoned Github - at least to the few people who seem interested in what I do.
Combine this with the use of StackOverflow Careers and Github Jobs, I might as well have died as far as potential employers are concerned.
Most employers know what a sudden and abrupt change in the graph really means. If someone's so short sighted enough to disqualify you on your contributions graph without so much as asking "why?", they don't respect your profession.
It is possible for another student at the Uni to find the repo, clone and turn it in and have us both up on plagiarism charges. It's unlikely, but I'd rather not run the marginal risk.
It was advised in my first year by a few tutors to just keep it in a private repo until hand-in. I'm fine with that, in fact, just publically listed a dodgy encryption algorithm I wrote for a class!
This could be abused and will render the whole thing pointless for everyone: make private repository, write bot to commit every day... now you have 4000 useless commits on your profile which an employer is supposed to think are legit.
I work for a company that does not uses github at all. We store code on our own servers. Do you seriously think that anyone working in such company is going to be unemployable?
I would say the companies you look at work in have seriously flawed hiring practices. They would not hired our best engineers, cause they do not have open source github contributions. They code plenty for work through.
If a flock of employers were requesting this feature it might be more reasonable. A candidate requesting this feature seems odd when the solution is obvious, and you know what the solution is already: write more code that you own.
It's different for every business but typically a CEO is the visionary, strategic leader of the company while the president leads the day-to-day operations.
If Github started selling insurance, there's a good chance they would have a president for Github Code and a president for Github Insurance. They wouldn't have two CEO's, though.
I know SC2 is probably a poor example but it fits very well here. It's the difference between managing a team strategy in an RTS versus looking at the individual unit deaths/spawns, types of units, etc.
The two are similar, and you'll often see them combined. For the most part, president pre-dates the CXO nomenclature. Before there were CEOs, COOs, and CFOs, there were Chairmen, Presidents, and Treasurers. Those titles/roles still exist, but are often merged with CXO functions.
While there are a lot of responsibilities that might vary company-to-company, the CEO reports to the Board of Directors (who represent the shareholders). Shareholders have control over the company because they can elect the Board of Directors (and depending on structure the CEO). The board has control because they get a vote on important matters like major officers and big expenditures. Depending on structure, they may elect the CEO. In the case of Github, it looks like the transition was driven by a friendly consensus of everyone involved. The CEO may need to get Board approval for changes in major officers, acquisitions, and major expenses, but they have the most control over those aspects, and thus the high-level corporate direction.
Because the CEO represents the shareholders, legally his or her main job is to protect the interests of the shareholders. For a public company, the CEO is the one sweating the most about the stock price; for a private company, the CEO is usually responsible for raising investment and driving the financial structure of the company. If someone is a potential investor or acquirer, they will be working with the CEO. Because they're selling the vision of why the company is a good long-term investment, the CEO tends to be the voice of the long term strategy.
The president may still have some of those responsibilities, but in most cases is less directly involved in fund-raising and more focused on the day-to-day operations.
That said, there are a number of other responsibilities around direction, operations, hiring, etc that can be split between the president and CEO.
The CEO most certainly does not represent the shareholders. The shareholders elect a board to represent them. The CEO represents the company's management, and is responsible for attaining the goals the board sets for him/her. It's worth pointing out that those goals aren't necessarily getting a high stock price or maximizing revenue. Lately, there have been a number of companies that set goals that have more to do with social goals than specific financial goals.
Sorry, it should have read: "Because the CEO represents the interests of the shareholders".
While the CEO is not (absent other roles) a representative of the shareholders (as you pointed out), there is a legal duty of loyalty to act in the best interest of the shareholders.
So while not a representative of the shareholders, the CEO is always responsible for representing the interests of the shareholders.
As you stated, those goals aren't necessarily related to the stock price or maximizing revenue. However, if the interest of the shareholders is something other than those goals, the CEO still has a legal fiduciary responsibility to act as a representative of the shareholder interests that do exist.
I suppose that's true in principle. But in practice, the CEO has to be pretty flagrantly acting against the shareholders' interests (embezzling, "cooking the books", etc) to be held legally liable. The case law here is predicated on the belief that "judges are not business experts", and I agree with that sentiment. Courts shouldn't be telling company management what is and isn't in their shareholders' best interests. It's up to corporate management to make that determination, and that's the way it should work.
Quite right it is the chairman of the board (or in this case the president) who is effectively the boss of the C levels and its the boards responsibilities to look after the share holders a CEO/COO/CFO is just another worker.
It's a good question, as it's a very vaguely defined job title, but generally the CEO sets overall vision, and managing the board, investors, etc (i.e. external facing) - and the President focuses on execution of that vision (i.e. internal facing). President/COO are interchangeable in this case.
Think about Facebook: Zuckerberg sets direction/vision, and Sandberg executes against that. (I'm sure it's not as clear-cut in reality, but you get the idea.)
In some larger companies, there are multiple presidents with responsibility for their own divisions, who all report up to the CEO.
That's really the question when it comes to having a separate president and CEO role (and often times having a separate CEO and chairman role). Who's really in charge, and what does the person who's not in charge do? There really isn't a good way to tell. Generally, it's either a sign that:
1. The President is being groomed for the CEO position once the "real" CEO leaves (such as Tim Cook).
2. One of the two needs to give up their position, but isn't because either nobody wants to force them out or they're too powerful to be forced out.
3. The company is so huge, the CEO needs to delegate some of their authority.
My money's on #2. Tom Preston-Warner clearly doesn't need training to be the CEO, and github clearly isn't big enough to need separate CEO and president roles.
Typically that's where Haley sits, who is typically the person you'll first meet when you're visiting or if you're dealing with some logistics with the SF office.
(As an aside, the idea behind the Oval Office as a visitor's reception is to turn the notion on its head a little bit. We've all been to companies where the reception area is just an afterthought, and you're stuck awkwardly standing in the middle of an open floor, not knowing what to do next. Our Oval is a tongue-in-cheek jab at that, and an attempt to make things a little more, well, interesting while you're waiting for a GitHubber to meet you.)
To add to what Holman said, at most companies the receptionist is considered unimportant, and is treated as such.
GitHub turns that on its head. Haley sits at the president's desk and is entrusted with people's first experience with GitHub's HQ. She's an important part of our culture. It's an big responsibility and a position deserving respect (as all positions are).
You're absolutely right. They say a picture is worth a thousand words - animated gifs even more so. I could almost imagine a world where animated gifs replace all kinds of communication. Press releases seem particularly ripe for change as they are usually wordsmithed to the point that any impression given is likely false.
We can easily expect GitHub to continue its exponential growth. The timing seems perfect to me...Chris will take on a role that will require him to ensure that the product remains excellent and will implement the innovation that Tom will be responsible for. Both suit their styles.
I love Github, but they need to build out more serious features into their issue tracker for it to be useful as a software estimation tool. Specifically I need:
GRANNT Chart: A Grantt chart is a type of bar chart that illustrates a project schedule. Grantt charts illustrate the start and finish dates of the terminal elements and summary elements of a project, such that completion of the project qualifies the team for a research and development grant.
Serious answer: Issues are not--to my knowledge--a software estimation tool. I have worked for decades in estimation-oriented development environments, and in my (anecdotal) experience, it is a mistake to use the same tool for both estimation and delivery.
It is seductive for them to be the same thing, as you can feed actual times back into your project, but the conclusion I've drawn is that the estimate reflects a kind of "interface" visible to the outside world, while the issues reflect an "implementation" visible to the team.
When they are the same thing, you end up doing a lot of work fiddling with the estimate to keep it congruent with the internal tasks, even when that congruence is not relevant to the delivery date(s).
Personally, I disagree. Sure, first you get time estimation. Then you get task-based time tracking, then you get average resolution time per committer, then you get weekly spreadsheet generation and analysis meetings on thursday afternoons and there you go, now you're getting memos about adding coversheets to your TPS reports.
I hope they are very careful about adding features and would rather see them focus on trimming less useful features.
Honestly I think that, like version control, time estimation can be a very useful feature for people working on a project alone if they use something like Joel Spolsky's evidence-based scheduling[1]. It would be nice if GitHub provided a way to keep track of that. I agree with you that there's a point where there get to be too many features, but if they're building a project for teams to work on things time estimation seems too useful to pass up.
The title is a formality, their organization is notoriously flat and being co-founders they probably don't make major initiatives without the other's approval anyways. If anything they're probably just swapping "who sits in the spotlight for this year".
And when you head a tech-oriented company, the difference between CTO and CEO is probably negligible anyways.
I think Tom and Chris are doing a fantastic job running and growing github. I'm just curious how this idea of trading CEO roles without making it public for a while works in practice. I think it could potentially matter to their employees, along the lines Ben Horowitz discussed in his post, "Shared Command" [1] - different situation, but analogous potential muddling of the chain of command.
Obviously some companies with both a CEO and a President can be amazingly successful [2], but tend to still be clear about who is ultimately in charge. It's not that I doubt Tom and Chris share a vision for github and clearly communicate and execute on that vision together; I just wonder when and how that model might hit its limits.
GitHub is not a company floating on some publicly-traded market yet. How they operate internally is of very little business to those on the outside. GitHub can operate anyway they like.
Did I say anything about Flash? I was talking about a video tag. And are you saying that a 4 seconds video in the video tag will be less efficient than an img tag with an animated GIF?
Personally I see no point in using GIFs, unless on a site which allows embedding images, but not videos. On a site which you control, video tag works just fine.
- ignore older browsers (today practically all browsers support the video tag, caring about older browsers in this context is like caring about IE6, i.e. it's bad).
- stick with a compromise solution (H.264, yeah that's sad, but until Daala will come this mess will remain, but now Firefox already supports H.264 easily, so it's not a problem anymore). If the space allows it, I wouldn't mind encoding in both WebM and MP4, just for the sake of supporting free formats.
I'm guessing for legacy reasons. They've been using gifs for a long time. If this was the first time they needed an animated video, perhaps they'd jump to video tag.
2014 should be a great year for them.
[1] https://github.com/about