Frankenplumbing
Wednesday, July 13, 2005, 09:46 AM - Tech
This from Mark Gibbs of Network World:My house is full of Frankenplumbing. There's a maze of copper, galvanized, brass and plastic pipe-work lurking under the house, and some of it is sound though much of it is decaying. And it is draped from one beam to the next and winds around and through joists. It has been cut off, capped off, welded, brazed, extended and joined.
The problem is if you do anything that disturbs the status quo of your Frankenplumbing your life will become much harder. Any minor change will usually break something, and major changes are guaranteed to cause chaos.
This then is the real foundation of IT, particularly at the enterprise level: The IT group is there to fix, add to and sometimes replace the Frankenplumbing for as little cost as, and with the least disturbance to the system, as possible.
That tells you something important about the kind of people you need in an IT operation. Among all the groups, the analysts, the programmers, the support techs and so on, you want to have a fair number of MacGyvers who are happy to grapple with the Frankenplumbing and capable of using some digital bailing wire and virtual chewing gum to create a workable fix.
Security
Wednesday, June 15, 2005, 02:42 PM - Tech
This article by Winn Schwartau sort of goes with my previous post about openness. I wrote about how to keep the data open from proprietary formats and controls. Schwartau writes about the basic rules on how to keep data safe.IBM Patents
Monday, April 11, 2005, 08:58 PM - Tech
The release by IBM of 500 patents could seed some fascinating innovative growth.First off, IBM could be doing the whole PC giveaway all over again. In the same way that the company published the architecture of the IBM PC in the 70s, IBM is giving away another chunk of intellectual property. Since I don't know the quality and timeliness of the patents, I'd been delighted to be educated by a kind reader.
Second off, it is, of course, a challenge to other big guns to divest themselves of patents. And, of course, the strength of the challenge depends on the quality and timeliness of IBM's patents.
In the first instance, if the IBM gambit plays out the same way as the PC publication, you'd hope that the company has learned and is prepared this time to maximize on the disruption.
In the second instance, the best outcome is that the cumulative pressure would allow Microsoft to open up Windows 95 (under the MSPL: MS Public License).
The worst thing that could happen is that the patents are essentially uninteresting and the effort thuds.
List of IBM patents
And here's something from Chris Nolan at eWeek.com
The bottom line: IBM has played some smart politics here. It has satisfied its obligations under open-source agreements, it has racked up brownie points with the open-source community, and it has created a kind of political confusion that will help it exploit its new patents. It's a clever strategy that just might work to the advantage of IBM and its competitors.
Format Openness
Friday, March 11, 2005, 08:20 PM - Tech
Speaking of openness, I just spent about two hours messing around with a spreadsheet back and forth between Excel and OpenOffice Calc. Enough to drive you crazy. It is a simple grid to track speakers at a meeting. It is meant to be printed out.I don't mind that one app doesn't support a feature in another app, but why does the one have to strip the formatting it doesn't understand (rhetorical question)?
Something's got to be done, and it will at some point. Railroad gauges were standardized. Electrical supplies were standardized -- outlets, plugs, light bulbs, etc. So too will formats and treatment of data become standardized.
I just hope it will happen sooner than later. The key is to overcome the Microsoft domination. If the computing environment remains a monoculture, then there will be no need. But as more applications become widely used and depended upon, then the economic irritation coming from the lack of portability will prove overwhelming.
Openness
Thursday, March 10, 2005, 08:38 PM - Politics, Tech
I just finished listening to this podcast from the Gillmor Gang (yep, just got an iPod, for that extra 20 minute walk to work). Towards the end, the Gang riffs on openness - open source, open standards, open formats, open economics, and open systems.This got me to thinking about priorities. The normal hierarchy of these opennesses seems to be:
1. Open systems: this is the utopia of complete interoperability with little or no external constraints, what the Gang talked about with the constraints of proprietary silos; akin to the radical democracy of one person, one vote without the influence of money and influence peddlers.
2. Open economics: this is the free beer ideal, where prices are directly tied to costs of production and their is no extra margin due to proprieatry control on processes.
3. Open standards: similar to a standard legal code, where the everyone plays by the same rules.
4. Open source: allow the free examination and modification of the code which drives the applications.
5. Open formats: keep the data free of constraints so it can be ported to any system without impediment. This extends to the metadata, too, such as the formatting data of a word processor.
Practically, in the management of an IT organization, my priorites run in other directions:
1. Open formats: the foundation of my job is to protect the data. If the data is constrained, then someone else has control over my data.
2. Open standards: to reduce the constraints in communication between applications.
3. Open source: this cuts two ways - reducing the cost of licensing and allowing examination of what the application is doing to my data.
4. Open systems: keep focusing on open interoperability of all components. This becomes a decision factor when all else seems equal - Red Hat or Suse?
5. Open economics: most consumers, both business and individual, are dealing with the same economic conditions, so the field, while not ideal, is relatively even. Sometimes that $150 proprietary app doesn't muck with my data and is certainly cheaper than having a developer toil for three days.
Almost-free Music
Wednesday, March 9, 2005, 01:51 PM - Economy, Tech
This suggestion makes a lot of sense to me. I'm not a music afficiando. I don't follow the trends. I don't listen to music that much.I buy music on a whim, and 15 bucks per whim is more that I'm willing to fork out. Even a buck per track makes me hesitate. Instead, I'm willing to spend the time scouring the web for legitimate free mp3s. But a nickel, even a dime, per track is something I'd seriously consider, particularly if there's the ability to stream a preview of the track and a link to the performer's site.
See, for me, scouring the web is in itself entertainment. If I'm sitting down for a couple of hours of mp3 searching, checking out the visuals and content of sites is as much fun as finding a good couple of free tracks. So the purchase threshold for me is very low.
Even the annual-fee, unlimited-download sites make me pause. Since my purchase barrier is about a dime, that means I need to buy about 250 or so tracks per year. I'm not going to make that kind of commitment for something that's a whim. It reminds me of the times I've signed up for gym memberships - great intentions that didn't pan out.
On top of it all, I'm offended with the current pricing because it has no relation to the cost of production. I don't buy the industry song-and-dance about the ancillary costs such as talent development, marketing, etc. and etc. I feel like I'm getting taken, particularly when I read that the artists still get squat out of the deal.
I think there's plenty of folks out there like me. Give us a price point that makes sense, and we'll fork over the money. Otherwise, we'll either a) keep scouring, b) download illegally, and-or c) ignore that market.
Truck Stopping
Wednesday, February 23, 2005, 03:15 PM - Tech
Roland has an interesting bit about truck-stopping technology. Trucks are mounted with a device which, when activated by remote-control, Developed by the Lawrence Livermore Natl Labs, the intended use is two-fold. Cops can stop trucks suspected of being hijacked by terrorists. Sensitive installations, such as chemical plants, can set up a perimeter that stops all trucks from entering. All well and good.
But let's look at the moral hazard. The chemical plants relax overall security because of the perimeter defense. That's not a good thing. And what about the terrorist who removes the activator before setting out to plow into a chemical plant? The perimeter defense is disabled.
Or the unintended consequences. Can you imagine the mess caused by a prankster on the LA freeways during rush hour?
Programming Planning
Monday, February 7, 2005, 06:45 PM - Tech
Dror Eyal has a good piece on how to survive in today's software market.At work, we do a great deal of our own development, and over the past couple of years, I've settled on the following as the proper way to develop smaller projects:
1. Describe the purpose of the application in less than two sentences.
2. Meet with affected departments and negotiate the scope of work and requirements.
3. Determine the interaction of the application to existing standards, both internal and external.
4. Mock-up user and administrator interfaces.
5. Define data flows and structures.
6. Mock-up programming flow and structures.
7. Code project; test; modify; release.
The design of an application should be modular. Here's another take, another way of conceptualizing it, from Roland Piquepaille.
Update: A more detailed and expanded write-up is here .
Next