Tuesday, August 10, 2004

Simple Systems: Effective Means

A topic sometimes debated on XML-Dev is that "worse is better", but the real topic is "simple is often better". Not always, but often. There are lots of examples from HTML (not so simple anymore), blogging formats, and further back, T-34 Soviet tanks, and not so far back, pagers. As one reads about 9/11 it quickly becomes apparent that the more sophisticated systems failed most frequently. The simpler systems like pagers were more effective but unfortunately, least distributed.

For those of us in the public safety industry, this is a lesson to be taken to heart. Command and control are worthless without communications and intelligence. While the dispatch center does thrive on sophisticated mapping and analysis technology, a command is most effective when it is presented in its simplest imperative form, such as, EVACUATE NOW!


Michael Champion said...

Just to resurrect our eternal xml-dev debate, I'm wondering if you have changed your mind, or that we've been talking past each other all these years. IMHO, "worse is better" in the original Richard Gabriel article basically means that "simpler for the implementer" is better overall than "simpler for the user." "The Right Thing", on the other hand, is the [usually forlorn] hope that making things simpler for the 'luser' will make the system work better overall. "Worse is better" is a great catch-phrase, but a lousy precis of the argument.

Having spent most of last night and this morning cleaning spyware/adware off the kids computer, I came to see one big reason why trying to make life easier for the end user often backfires: You're also making life easy for the real losers who write slimeware.

Speaking of the T-34, I guess Metcalfe's Law is a variant of Stalin's dictum that "quantity has a quality all of its own." The sheer quantity of HTML, RSS, pagers, etc. out there creates a lot of value, irrespective of the quality of the technology in technological terms.

len said...

Sometimes I'll take either side in a debate to see to it the debate takes up all of the issues. Not surprisingly, that position is misunderstood in communities that insist on adherence to the party position. The joy of blogging is that I can relate my reactions closer to real time. You are right that 'worse is better' is a rotten way to start the debate, but where the energy to create logic stems from the emotional, it is a way to start the fight.

1. Simple as possible. No logistics expert argues with that one. More moving parts means more fault modes means a smaller Mean Time Between Failure.

2. "As possible" means that one analyzes the device in the context of the mission and ensures that the failure modes of the device match the potential situations one will apply it in.

3. The web as 'the scalable universe' leads to one believing that if it doesn't scale to the web it isn't worth working on. This ignores that there are scales of application that are not isomorphic to the web. Since the web architecture has the conceit that it is isomorphic to the universe, and given Boltzman entropy, the web device can become simpler than is useful for some applications. In other situtations, both the public internet and the web device perfectly match the requirement. Caveat emptor should prevail but the reality is caveat vendor. XML is a product of simplifying SGML in light of lessons learned on the web. In most cases, it is the ideal messaging format but not all.

A cell phone is a great way to communicate when one is mobile and the situation is relatively benign. On the other hand, as someone who recently had a car in the shop for three weeks because a benign mammal with a cell phone tucked neatly under her ear rammed it from the rear, some situations are not so benign. The situation on 9/11 crushed first responders between their major incident plans and the unanticipated results of jet fuel burning furiously against metal 1000 feet in the air, and others in a situation where the dispatch centers were coping with unreliable information plus the urgency of their location.

Overbuilt systems tend to fail in the face of catastrophes. Even WWW systems. Even simpler systems such as RF.

The problem of catastrophe in formal systems have been studied over the years in various applications from biological systems, plantetary systems, to command and control systems, and so on using varying theories of chaos and complexity. Most of these can be boiled down to situated networks of communicating nodes. As the frequency of exchange and the complexity of given messages rises, so does the failure rate. Catastrophes are full of unknown unknowns. It is difficult to create major incident plans for such. Major incident plans rely on the ubiquity of reliable assets among these, communication devices. The rule of thumb seems to be that in such situations, simpler systems are more reliable. Selah.

Unfortunately, the procurement cycle precedes the fielding of assets and it is often complicated by consultancy and the desire for assets that confer status. For that reason, logistics analysts are used extensively by the military and should be by public safety procurement consultants.

Web technology receives a lot of press. That makes it a status conferring technology. But in some cases, it isn't the best solution or effective. Platform vendors must take on the role of logistics analysts and step up to the job of leading their industry instead of only reacting to RFPs.

len said...

BTW, Mike, I should add this: the art of analysis for design should include the notion that it is very important to identify where or in which components simplification benefits any user. Simplification for the developer can result in complexification for the end user. A common result of thin client design is that the user has to wade through many pages and forms to complete a complex transaction that could be better organized for the end user by a fat client with more local state management.

Then the problem that the user model is another 'no size fits all' challenge is to be accounted for. Jakob Nielsen might not be the best guide here in all cases.

Scott Edwards said...

Are you tired of getting nice compliments on your blog, when really you're not makeng the kind of money that you deserve? Now you can join a completely FREE program. No buying and No selling. FREE to join. All you do is refer customers. To get started fast, click here: based business home internet opportunity work site. It pretty much covers based business home internet opportunity work related stuff and it's FREE to join.

Scott A. Edwards said...

Here it is... FREE advertising, FREE download. No cost to you! Get your FREE download NOW! Make money and get FREE advertising! This is a great program for the new year... Check this out now for FREE!

To find out more visit: hi-tech site. It successfully exposes FREE information covering traffic and hi-tech related stuff. Don't forget, FREE, FREE, FREE!!!

Extra said...

Hi blogger:)

Well done - I can hardly receive a comment on my blog. The only comments I have are ones that my brother created:) Hope you keep yours good as much as it is now! Cheers:)

Remove Adware

Online said...

Hi blogger:)

I shall inform all my school friends about this blog... they were searching for the same information a couple of days/weeks ago. If they still need that information they will adore your blog:)

blocker spyware

Comment Policy

If you don't sign it, I won't post it. To quote an ancient source: "All your private property is target for your enemy. And your enemy is me."