If you could have a datacenter/lab/systems management tool that doesn't exist built for you (whoever you happen to be), what would it do?

Not just in terms of adding new features on an old one, but something entirely new and different?

I think we're a bit self-limited with the current model of Install/Config-Manage/Monitor (it's what's always been done) and am wondering what everyone's thoughts of the future of controlling large numbers of machines are. Make no assumptions, impose no limits (or PHB requirements), how would you want things to be to really shake things up? What still sucks today? Where could you go if you had technology to enable it? Is there a tool or program (or artificial intelligence [1]) that doesn't exist that you would like to use and possibly work on in the future? Something to join the modular Cobbler, Func, Puppet/cfengine/bcfg2/other family or maybe to go in a completely different direction? Ideally this would be something that could be started with a small prototype and grow over time with contributions as others see fit?

----

Footnotes:

1. The Matrix on Blu-Ray is rather kick-ass and makes me disappointed with my lack of a neural interface all over again. [2]
2. I don't know Kung Fu, does Python count?

I've been thinking lately about what the positive impact of what I do is, beyond code. Ideally I think I want to be about encouraging cross-organizational collaboration to build new exciting and unpredictable things, and I think this is what OSS is really about (but not everyone realizes this). That all being said, part of the ways we encourage collaboration is not just in creating collaborative software projects, but spreading the message about what collaboration and free exchange of ideas can do. As much as said talk doesn't do much, it's not quite true. We can easily start and run a few projects ourselves -- but we can /talk/ a lot. Communication is important, and we can show people what we do with the few examples we have. So, yeah, collaboration, ideas...

I think, right now, we are NOT getting that message out to the academic sector enough. It's not just about software development methodology, but really it's a game changer -- if you want to call it the bazaar, fine, it's the bazaar. So, what am I getting at? We need to be encouraging new interactions that generate new, truly innovative things. Right now, I think a lot of what we see everywhere is still evolutionary, and small micro improvements on projects. An installer is still an installer, a desktop is still largely a desktop doing the same thing a desktop always did, a web browser is still a web browser. Zzzzz? We keep rewriting things, and that's not bad -- but the whole world is also rewriting things. I feel myself getting into this rut. What we really want is a LOT of new ideas on the table and to see what sticks. Not just ten of fifty ideas, but a thousand. Ten thousand. Colleges are a great places for this to happen in addition to industry (often industry has more pressures to release products and less time to try new things), but there are some things holding colleges back. For one, folks aren't encouraged to build longstanding projects that live beyond the assignments (or involve students across semesters) -- two, universities often try to own IP for things done on machines owned by the university. Both are not creating this huge interplay of ideas that I'd like to see. (Ideas breed with one another, genetic diversity of ideas is what we're going for).

What spawned this -- Slashdot had a blurb about how student inventions are increasingly being patented (for license fees) by universities. The Link says something about licensing an invention back to a student for $75k upfront + future royalties. What's the likelihood that the student can afford this? Are we building collaborative, nation-transforming innovative industries this way? (Moving from large industry to garage-startups?) Or are we just training replacable workers and not designers? What ideas are we producing that can multiply with each other?

When I was back at State (by which I mean NCSU) several of us (students) had the same theory about this holding innovation back -- in the light of NCSU not getting royalties from SAS (the world's largest private software company and also the largest software company I know of that has a Yoda statue in their lobby), they are discouraging students from doing interesting things by owning rights to what they create. As a result, innovation is reduced. We didn't see why anyone would start a Google now, and as a result we were very leery of innovating inside of our class projects. We didn't, and I am pretty sure if we had been allowed to think about building businesses while we were around a huge crop of smart people we would have done some amazing things. Imagine if instead of busywork class assignments we were building new technology that people could really use. What if people were told instead of "here's an assignment to build an blackjack simulator" you instead get "here's your mission: find a way to solve the world's problems WRT ______".

Perhaps Universities should realize that folks like SAS give a LOT to the unversity in terms of donations, and also hire their graduates, which feeds the universities more than the patent royalties. As a result, while NCSU may be able to extract some revenue from a few startups, it also makes sure that innovation that can start outstanding companies and products is curtailed. There is also lost affiliation with the brand of the university -- for instance, everybody knows Google came out of Stanford. Heavy royalties probably would have prevented it's formation -- or even the formation of the idea.

The open source way of doing things would allow for free exchange of information and would understand that increasing the local economy and encouraging new ideas to flourish is more advantageous than controlling ideas and IP. Imagine the RTP economy without SAS, for instance. Now imagine three more SAS's that you don't have today, each with lots of perspective alumni donors working there and collaborating regularly with students. Which would you rather have? Which are you building?

Another point: funding. I had a LOT of time in college to learn things and do assignments. Most good OSS work today comes from paid developers working on a project of some kind or another for their company. One yields advantages people can share, but the university side is seemingly untapped aside from a few shining examples. In general, code developed is thrown away and not seen/used again, and usually it's just reimplementations of something someone did the previous semester. Everyone time to come up with crazy ideas and implement them then. How many did? Was it because they weren't told they could and the framework was just conceptually broken?

Thankfully I work where I can implement new ideas that I come up with -- but generally, for a lot of people, they usually have to implement ideas come up with by other people. For most folks, universities are were folks will have the most freedom. Let's encourage them to innovate and keep the rights to what they create, and trust that they'll do the right thing later. We might even have our flying cars by now.

Wouldn't it be amazing if instead of training tomorrow's coding labor force, we built 10,000 startups and incubate new ideas? Are there more ways we can spread OSS and this thought process (which seem naturally linked) amongst our educational system? Apple started out of a garage. The next game changer isn't going to come out of Redmond, either. Turn the kids free and see what they produce.

Of course, please work on the Three Laws before Skynet. Thanks :)

Unfortunately, this blog is just words -- we need to find new ways to create more decentralized communities for people to collaborate and share brainpower. How can we use what we know from things like Fedora to enable more people to freely create and share? And how to we tap universities and such when they are young and before they go to places that don't have the freedom to encourage as much innovation? Any ideas? We need to quadruple our efforts.

We're looking at Obama promises such as open government and moving things to be more democratic again. This is great. There's potentially a new wave of funding to creating green technology (also great). At the same time, we need to make software and invention more democratic and accessible to everyone. That would be a great role for all of us to help foster. The potential is a technological renaissance. I'm not in academics, but I see huge potential for their taking OSS (and the mentality behind it) to the next level. It's time.

We need our educational institutions to help that happen, and we have to help them see that potential.

Maredsous 10, a nice Belgian Tripel, with incomprehensible things written in Belgish (1) on the glass. Yum.

And another thumbs up for Tyler's Tap Room's Garlic Fries.

---
footnotes:

1. It is too a word! (2)
2. Apparently not made my monks though, darn.

And the weirdest thing to happen to me in 2008 happens on ... December 31st, 2008.

I'm walking through Bond Park (it was a bit too cold to hit the local MTB trails and I was lazy) with the camera, and I come upon three people walking in the distance.

I look, and realize the first person is apparently wearing a very furry parka. No, it's not. It's ... a bigfoot or gorilla. And the girl is walking it on a leash. Ok, maybe it was a kid in a suit. It was slightly bouncy, but generally oblivious to my presence; perhaps it lost a bet.

The fact that it's out of focus is highly ironic; after all bigfoot photos are almost always out of focus -- but I swear, the AF just locked on a tree; I didn't mean to.

Believe in bigfoot. Believe.

I did get to go to Hawaii though I only uploaded 100 or so pictures.

Best of 2008 Slideshow Link.

Next year: take (and upload) lots more photos.

I thought this post by Michael Meeks, (see also slashdot, where I found it), was outstanding. I've written about similar things before -- in this case, they have some graphs to back it up. (Never mind this only tracks /commits/, which is why, eventually I think we need much more intelligent community tracking software -- EKG does not yet deliver)).

Projects are open to the extent that their upstreams are welcoming and diverse. It is for this reason that simply "open sourcing" an app is not sufficient. Similarly, I don't see much value in simply Open Sourcing Domino. Dumping code and calling something OSS is not enough. It will remain crufty and may never become a great space for people to collaborate. You have to constantly look at building and rebuilding your project into a space for collaboration. Remember, it's not just about the final product, it's about the journey.

Communities have to be grown, tended, stroked, stoked, lead and fed. You can't do it in a box, you might be able to do it with a fox.

The ultimate challenge is really to correlate community openness and success with application success.

Currently I would speculate that more-closed communities have applications that are (A) more stagnant, (B) harder to use, and (C) much less feature rich.

If we apply this logic to OO, well, I really don't see any innovation there. It's basically the same program it has been since the first release of Star Office. Applied to Firefox, I think we see the same things -- it's basically the same browser, and the reasons for "shrinking" the code (when Mozilla spawned Firefox) need to be done again. Firefox needs to be more stable and zippy once again, and essentially it too has been the same browser for many releases. Where are the new features and innovations?

(At the ultimate closedness levels, you have Windows of course -- which has been, fundamentally the same user experience since Windows 95.)

Can we take from the article any reverse-metrics? If you see your project stagnating, ask why. It might be that you're not opening yourself up to outside ideas, or aren't doing enough to lower barriers to contribution. Once things get past a certain stage, creating environments to collaborate is more important than creating the software.

So, if someone Open Sources Domino, what happens? I think they get some free users. They'll probably get a few patches, but I have this feeling it's going to be this scary, tear your head off codebase, with an upstream that doesn't know how to run it. I hope they do, I think many of us would be more than glad to show them the way, but it's not something I'd hold my breath for.

In terms of OSS apps, we have a lot of huge collaborative successes. We have some decent successes that succeed because they have been well funded by corporations. Ultimately the latter can become the former, but they have to work at it.

I think there's something to be said for the domain too -- you typically don't /want/ to work on a word processor unless that's part of your job. The same goes for web browsers. It's much easier to justify investing in a tool when that tool is directly a part of your job and making those enhancements can lead to further successes down the road. In that vein, perhaps we shouldn't be so quick to judge things like OO, but instead think, "if I could run this project differently, and make it simpler or more inviting, how would I do so?"

One might start by including a First Person Shooter Engine. Seriously, I don't know.

Do this though, regardless of what you work on -- look at your community mix now. If you don't like it, do /something/ to work on it. It doesn't matter what, just make that something you look at every day and try to keep improving. If you do like it, think about why you like it. Keep asking questions and share your results.

The idea that OSS just magically works better than other software development methods is a myth. You do have to work at it. /Then/ it works fantastically better.

This one is a boring blog memes compared to the zombie one, but I'll bite:

1) Which was your Linux distribution story?

First distro? NCSU's Realm distribution counts, right (even though I favored our Solaris machines then because they were a bit more stable?). The first thing I ran at home was Debian, though I forget the Toy Story versions -- it was "testing" or "unstable" most of the time. Liked it a lot, including how nearly everything was packaged already. At my first work, Red Hat 6.2+ and lots of Debian mostly. I think I tried Fedora 1 at some point, but my first serious usage was 2/3, primarily as a development platform for embedded Linux work at a startup. I also have used lots of distros that aren't around anymore, and a fair amount of SuSE -- but I don't like it based on all the tweaks they made. The United Linux stuff, the bomb background, stuff like that... Since picking up FC3, I haven't really run Debian since. Obviously some EL is thrown in there.

2) What is your preferred $your_distribution version?

I am pretty happy with Fedora 9 + KVM,. Fedora 10 still has some bugs and some nits that make my life doing install tech/testing a bit more "fun", though it's our job to report those and make it better, right? They'll get hammered out. My non-VT capable laptop looks forward to having Xen going again someday once I get around to upgrading it.

3) Write a short story (more like an anecdote) about your past distribution.

I wish more things had super cow powers, even though there are very few super cows. Super sheep I do know about from Worms and they are VERY dangerous.

The final stretches of the race in "Herbie Goes to Monte-Carlo" bears a remarkable similarity to the track in Gran Turismo 4. I was watching it and thinking "here comes the switchback", and it was right on schedule. Also the waterfront finish and the final tunnel.

I remember clearing it in a Lotus Elise, but a #53 Volkswagen Beetle would be a great addition for GT 5. I hope someone is working on that.

Recently I recieved a flickr email from some local business saying they wanted to use one of my photos (of NCSU) for their web site -- an area apartment. I wrote back and explained to them what "NonCommercial" was on the CC license and offered to license it to them for a small cost (non-commercial usage of my stuff is marked free if attribution is supplied). They replied back with "what was the fee?", and I replied "it depends on usage, what do you normally offer?". I was expecting a couple of hundred bucks as it sounded they wanted to use it in relatively large format, but didn't ask for any price, I would have probably taken less. The final reply was "We were looking for free image usage". Well, then, that kind of wasted my time and theirs, didn't it?

Ultimately when someone is going to use a photo for /advertising/ purposes, that photo is going to earn them a decent amount of money. It only makes sense that the photographer (who bought the camera, spent the gas, etc) is given some small share of that revenue for helping earn that money -- his equipment probably costs more than the computer that is building and/or hosting the website.

Preying on newbies by asking for free content is not a good way to do business if you're in advertising. It's cheap and just makes your company look bad. It will also probably result in you not getting your best choice of content, either.

So I'm sitting here playing with the synth and I've got it tuned to the most detuned aggressive bass sound possible, with a Boss Metal Zone in the insert loop, and lots of reverb.

Christmas Songs sound really cool this way.