Comments on: Quit Today http://digitalvampire.org/blog/index.php/2008/10/29/quit-today/ Linux hacker, recovering mathematician, former athlete Wed, 25 Aug 2010 23:50:28 +0000 hourly 1 http://wordpress.org/?v=3.0.1 By: Fab http://digitalvampire.org/blog/index.php/2008/10/29/quit-today/comment-page-1/#comment-856 Fab Thu, 23 Jul 2009 17:32:11 +0000 http://digitalvampire.org/blog/?p=50#comment-856 Hey Roland, I was just reading your blog, kinda randomly. Yes, the post is pretty ancient, but I want to point out a very important issue you raise, that of employer loyalty. It amazes me the number of people who expect(ed) their employer to have compassion and feelings towards them, just because they do(did) for the employer, and are completely blind sided when the numbers don't add up anymore and they're given their two weeks notice. Hey Roland, I was just reading your blog, kinda randomly. Yes, the post is pretty ancient, but I want to point out a very important issue you raise, that of employer loyalty. It amazes me the number of people who expect(ed) their employer to have compassion and feelings towards them, just because they do(did) for the employer, and are completely blind sided when the numbers don’t add up anymore and they’re given their two weeks notice.

]]> By: Roland’s Blog » Blog Archive » The enemy within http://digitalvampire.org/blog/index.php/2008/10/29/quit-today/comment-page-1/#comment-590 Roland’s Blog » Blog Archive » The enemy within Wed, 05 Nov 2008 19:20:30 +0000 http://digitalvampire.org/blog/?p=50#comment-590 [...] Roland’s Blog Linux hacker, recovering mathematician, former athlete « Quit Today [...] [...] Roland’s Blog Linux hacker, recovering mathematician, former athlete « Quit Today [...]

]]> By: AnonymousB http://digitalvampire.org/blog/index.php/2008/10/29/quit-today/comment-page-1/#comment-588 AnonymousB Fri, 31 Oct 2008 22:57:30 +0000 http://digitalvampire.org/blog/?p=50#comment-588 I have a similar problem to Anonymous. I've written, on my own time, an experimental improvement to 'git bisect' (not actually integrated into git) which knows about the case where the test has false negatives (because the bug is intermittent). But, my employer thinks they own everything I do (even though that seems to be invalid in my legal jurisdiction). I talked to my manager about this, and the result was not encouraging: even though she thought the company had no interest in the code, releasing it would apparently 'ring alarm bells'. The best suggestion she had was to write an academic paper (which would apparently fall into the category of acceptable professional development) and release the code with that. But I'm a bit stuck on the paper. Ideally I would release my tool now, and see if anybody thought it was worth pursuing, and maybe had some bug which they could try it out on and give me feedback. It's much harder to try and work it up to a publishable level when I don't even know if it's actually of use to anyone. I have a similar problem to Anonymous. I’ve written, on my own time, an experimental improvement to ‘git bisect’ (not actually integrated into git) which knows about the case where the test has false negatives (because the bug is intermittent). But, my employer thinks they own everything I do (even though that seems to be invalid in my legal jurisdiction). I talked to my manager about this, and the result was not encouraging: even though she thought the company had no interest in the code, releasing it would apparently ‘ring alarm bells’.

The best suggestion she had was to write an academic paper (which would apparently fall into the category of acceptable professional development) and release the code with that. But I’m a bit stuck on the paper. Ideally I would release my tool now, and see if anybody thought it was worth pursuing, and maybe had some bug which they could try it out on and give me feedback. It’s much harder to try and work it up to a publishable level when I don’t even know if it’s actually of use to anyone.

]]> By: Jamie Bennett http://digitalvampire.org/blog/index.php/2008/10/29/quit-today/comment-page-1/#comment-587 Jamie Bennett Thu, 30 Oct 2008 09:43:56 +0000 http://digitalvampire.org/blog/?p=50#comment-587 Changing jobs is a daunting prospect at the best of times, in the current economic climate even more so. May of us would love to contribute to open source software on a full time basis but instead get relegated to doing it in our spare time which isn't ideal. The best outcome for the majority then, as suggested, is to eat away at the inside; suggest open source to your company, actively promote it and be knowledgeable enough that when the company has questions, you can give good answers. Changing jobs is a daunting prospect at the best of times, in the current economic climate even more so. May of us would love to contribute to open source software on a full time basis but instead get relegated to doing it in our spare time which isn’t ideal. The best outcome for the majority then, as suggested, is to eat away at the inside; suggest open source to your company, actively promote it and be knowledgeable enough that when the company has questions, you can give good answers.

]]> By: roland http://digitalvampire.org/blog/index.php/2008/10/29/quit-today/comment-page-1/#comment-586 roland Wed, 29 Oct 2008 21:45:32 +0000 http://digitalvampire.org/blog/?p=50#comment-586 @Anonymous: Your situation is exactly what I had in mind when I said that Greg's advice to quit doesn't need to be taken literally. You have a job that you like right now, but the restriction on kernel contribution chafes. If I were you, I would keep pushing on the ban on any open source kernel work; it's good for your employer and good for you if you are able to contribute to parts of the kernel that have no conflict with your business. Just by being a kernel developer, you must run into kernel bugs completely unrelated to your core work that you are able to fix. For example, some of my commits such as 4c41d3ad ("ACPI: Always return valid 'status' from acpi_battery_get_property()") or bb10ed09 ("sched: fix wait_for_completion_timeout() spurious failure under heavy load") have nothing to do with InfiniBand or RDMA. It's just that when you run a bleeding edge kernel and you know how to debug the kernel, you end up answering questions like "why does my laptop crash on boot with two batteries installed?" With that said I think it's not a bad idea for you to quietly explore what open source kernel jobs you can get. If you find something that you like better than your current job, then you can take it. There are certainly plenty of companies hungry for Linux kernel developers, and since you never have more leverage than when a company is trying to hire you away from your current job, you should be able to negotiate contributing publicly (or just use the job offer to get concessions at your current job). I definitely don't think it's necessary to be a "rock start" to get hired. Certainly my work as a subsystem maintainer gets me noticed by a lot of peope, but my experience with the kernel community has been that it actually is meritocratic. No one had any idea who I was when I showed up with an InfiniBand subsystem, and while it took a lot of work, it did get merged when it was ready. Certainly having no public track record makes finding job opportunities more of a challenge, but that just underscores why it's in your interest to build that track record. @Anonymous: Your situation is exactly what I had in mind when I said that Greg’s advice to quit doesn’t need to be taken literally. You have a job that you like right now, but the restriction on kernel contribution chafes. If I were you, I would keep pushing on the ban on any open source kernel work; it’s good for your employer and good for you if you are able to contribute to parts of the kernel that have no conflict with your business.

Just by being a kernel developer, you must run into kernel bugs completely unrelated to your core work that you are able to fix. For example, some of my commits such as 4c41d3ad (“ACPI: Always return valid ‘status’ from acpi_battery_get_property()”) or bb10ed09 (“sched: fix wait_for_completion_timeout() spurious failure under heavy load”) have nothing to do with InfiniBand or RDMA. It’s just that when you run a bleeding edge kernel and you know how to debug the kernel, you end up answering questions like “why does my laptop crash on boot with two batteries installed?”

With that said I think it’s not a bad idea for you to quietly explore what open source kernel jobs you can get. If you find something that you like better than your current job, then you can take it. There are certainly plenty of companies hungry for Linux kernel developers, and since you never have more leverage than when a company is trying to hire you away from your current job, you should be able to negotiate contributing publicly (or just use the job offer to get concessions at your current job).

I definitely don’t think it’s necessary to be a “rock start” to get hired. Certainly my work as a subsystem maintainer gets me noticed by a lot of peope, but my experience with the kernel community has been that it actually is meritocratic. No one had any idea who I was when I showed up with an InfiniBand subsystem, and while it took a lot of work, it did get merged when it was ready. Certainly having no public track record makes finding job opportunities more of a challenge, but that just underscores why it’s in your interest to build that track record.

]]> By: Anonymous http://digitalvampire.org/blog/index.php/2008/10/29/quit-today/comment-page-1/#comment-585 Anonymous Wed, 29 Oct 2008 19:33:31 +0000 http://digitalvampire.org/blog/?p=50#comment-585 I work for a company that develops proprietary Linux drivers (no we do not produce graphics cards). Because of this I am not allowed to do any open source work on the Kernel even in my free time. When I complained I was granted permission to work on one open source project that was completely unrelated to anything our company does. This is fine, but I would greatly prefer to work on the kernel. So why do I stay? Well, I'm a key technical contributor to our Linux strategy and I see hope that I may be able to convince management that open source drivers will benefit the company. Additionally I'm a little scared that it will be difficult to find a job that allows me to contribute to the kernel as I am not an existing Linux kernel "rock star". I work for a company that develops proprietary Linux drivers (no we do not produce graphics cards). Because of this I am not allowed to do any open source work on the Kernel even in my free time. When I complained I was granted permission to work on one open source project that was completely unrelated to anything our company does. This is fine, but I would greatly prefer to work on the kernel. So why do I stay? Well, I’m a key technical contributor to our Linux strategy and I see hope that I may be able to convince management that open source drivers will benefit the company. Additionally I’m a little scared that it will be difficult to find a job that allows me to contribute to the kernel as I am not an existing Linux kernel “rock star”.

]]>