I've been a software developer for 6 or 7 years professionally. As a software developer, I have learned that I am not a number of other things. I am not a database administrator. I am not a release control manager. I am not a project manager.
Adam is a very brilliant person. He earned a 4.0 in Computer Science at Carnegie-Mellon and is studying for his Ph.D in Computer Science at Berkeley. I think something we need to realize is that we all have our strengths and weaknesses. Just because someone is a good programmer does not mean that they are good at the satellite tasks that I listed above. Adam and Nathan are really bearing a huge brunt of the responsibilities of this cooperative. I would say they are being overworked in fact. With overwork comes oversight. Not only that, but like anyone else, they have areas that they are both well-versed in, but others where it would be best to find some different support people, but unfortunately there are none. As FGB said, we are a cooperative. This is not my first cooperative. I have seen many lovey-dubby touchy-feely we-are-all-one organizations go under. I have seen others that refused to go under and kept operating at a negative, while most members jumped ship. Everyone was a nice and dedicated person, but they got slaughtered by the corporate forces in the same sector. When I look at the discussion of our problem from my perspective as a software developer, I hate to be brutally direct,. but I see issues in project management and release control, each of which I will go into. Having issues in these areas is serious. How many corporations do you think go in the red when they have to scale out? How many people have their jobs at the end of the day if they keep sucking funds out the company on a project that does not materialize? Project Management - is the ability to specify a project, corral the needed resources, and execute the spec within a specified time frame. We can look at the Perl 6 project (or the GNU HURD) as a prime example of a bunch of well-intentioned people that were great at software development, but had no clue of how to keep the show running smoothly in the meantime. Owing someone 4000.00 and then asking us about it with no clear ETA on project completion would get a _project manager_ booted out of any company anywhere in any country as fast as the CEO could get up and line up his shoe on the person's ass :) But again, we need to see this situation for what it is. We have a tremendously talented and tremendously compassionate and tremendously dedicated leader in Adam. We have the best software system for unix-based virtual hosting out there. But asking him to do all this and have clear vision as a project manager is too much. Many companies our size hire 2 or 3 full-time project managers. Again, this is not an exercise in finger-pointing, but if we provided an objective recount of what has happened so far to any IT professional, their curt assessment of the situation would be: "so you failed to handle the project management aspects of your project"... end of story. Release Control ============ Release control is the process of incrementally evolving a piece of software in response to business demands. I am not completely clear on why this migration is taking so long. I have heard some things about AFS issues. I am also not clear on why it is happening the way it is - migration to a different (improved) setup as opposed to raw-mirror and then improve. Again, taking the corporate metaphor to heart, it is rare that a company will shut down or incur huge operating expenses just because they want to roll out a bigger better product. Wal-mart is an exception. When their online website s00ked, they shut the whole damn thing down, contacted Sapient, and had it done right, and then re-started things. Sure they lost some revenue and some face, but they are past that now on the right track. >From what I can tell, it looks like we are trying to do 2 things at once. We are trying to migrate and improve. Either task is bug-prone and trying to do both simultaneously increases problems exponentially, not linearly. Migration is just that - moving our shit from where it is to a new machine :) Improvement is things like AFS, domtool2, etc. While I understand the Wal-mart/Perl6 mentality of "everything is fresh, shiny, clean and new now!" It is clear that a few voices have spoken up in displeasure at the increase in operating expenses. Not only that, but fyodor is not wal-mart: fyodor works well for everyone using it on a daily basis. We dont have something broken like wal-mart. Was the strategy for migration put before us as a coop? Were the project management and release control issues placed before us as a coop? What can we do as a coop to have better project management and release control? Who can we recruit that _is_ focused on money. It is nice to be a Justin Leitgeb - all heart and passion. Or an Adam Chlipala, all intellect and technical skill. Or a Nathan Kennedy, all legal and overworked. But it is clear we need a few more people with their dedication level but in the two domains in which this project has failed - project management and release control. I hate to draw analogies to the Iraq war, but what are we doing over there? Do we really expect the violence to stop? It is obvious that situation can only get worse and and can only suck more and more funds. I think we all need to stop and reflect on what skills we can offer to the cooperative to prevent situations like this from happening again. Let's network with other similar organizations. Let's get some communication going on what we plan to do before we do it. Let the members have some say-so in how these things go. We need a skilled project manager and release control expert to step up to the plate and harness our raw technical prowess. Technical prowess has only led to business success in the case of google. More often, it is business prowess and technical amateurs that net the big bux. We need some tight-fisted people with business sense now. We need project management charts and regular updates on progress. We need incremental release control after a mirror of fyodor on Peer1 resources. I dont like this Perl6 approach to creating a new tribe on a different island. Alternatively, what stops us from "bucket-hashing" coop members? Instead of one big new hosting setup, why not have fyodor-power-user.hcoop.net and fyodor-trivial-user.hcoop and bucket-hash people onto servers? Why grow out horizontally to accomodate the JC Hallgrens with the Terrence Brannons in one bucket when our needs are clearly different? I know this letter is going to aggravate some people. But I have something to say and I'm not going to let some hate by others stop me. Also, I have to congratulate myself. You should have seen the hate-laced finger-pointing email I wrote up all yesterday on this. It is amazing what a good night's sleep will do to one's perspective. On 4/28/07, Adam Chlipala <[EMAIL PROTECTED]> wrote: > To recap: Peer 1 has been charging us for colocation services for a few > months now, as we've been setting up our new servers there. We're not > yet ready to switch over to the new systems, though. I have been > working under the assumption that a reasonable slice of the membership > would be unwilling to pay shares of the Peer 1 costs before they are yet > receiving direct benefit from the new servers. As a result, I haven't > been incorporating the Peer 1 costs into the financial records on the > portal, the idea being that, at the point when we are ready to move any > member over, we consider all the Peer 1 costs to that point as a > one-time cost to be split in a special way to be determined. > > Now, a few months into that, we have about $4000 worth of those > charges. I sent a request to hcoop-announce recently asking for people > to throw in more money to help cover these costs in the mean time. With > the volume of people who seemed to respond to that, it's clear that only > payments of several hundred dollars per person could make a difference, > and that didn't happen. I've ended up essentially paying all Peer 1 > costs from my personal funds each month. > > So, what should we do? I don't have any idea how soon we'll be ready to > switch to the new servers. Essentially no one responded to my call to > "lend" HCoop money in the interim. > > Should we call this the point where we bundle the charges so far into > one sum that we split among willing members ASAP, and from here on bill > monthly charges to everyone as they arrive? > > Or, can more of you add several hundred dollars each to your HCoop balances? > > _______________________________________________ > HCoop-Discuss mailing list > [email protected] > http://hcoop.net/cgi-bin/mailman/listinfo/hcoop-discuss > > -- HIV/AIDS on trial - http://www.aliveandwell.org/ Let's play chess: http://www.SlowChess.com _______________________________________________ HCoop-Discuss mailing list [email protected] http://hcoop.net/cgi-bin/mailman/listinfo/hcoop-discuss
