Are you building a Sherman tank ?

[Folbec]

An interesting analogy here :

Of software and sherman tanks

I don't know if the article needs subscription or not, so I add the quote :

believe that in the swath of the software industry focused on business and consumer applications, we ought to build Sherman tanks. During the Second World War, German Tiger and Panther tanks had stronger armor and more firepower than American Shermans, but were difficult to produce and maintain. They had complicated parts and, when they broke down, often could not be fixed on the front lines. In contrast, the Sherman had thin armor, little firepower, and a high profile (making it an easier target to hit). It took about four Shermans working together to take down a Tiger tank. However, when they broke, they were easy to fix. They used fewer components than German tanks, and the components were easy to obtain and reuse.

The Sherman could be adapted easily to different tasks, and scores of variants of the tanks were produced—often modified on the front lines (Sherman would be a fine name for a scripting language)—to clear hedgerows, clear minefields, act as bulldozers, fire flamethrowers, travel amphibiously, and lay bridges. Sherman tank software development aims to produce resilient, reusable, adaptable, and easy-to-maintain software. The software isn't perfect, but it's easy to fix when it breaks and to extend to meet unanticipated needs.

The first step toward achieving these seemingly unattainable goals is to accept imperfection. Imperfect software makes it to market and gets used widely (Java is a case in point). Perfect software—or, more accurately, almost perfect software—rarely sees the light of day or ever gets used (does anyone remember Taligent?). It's not that (almost) perfect software isn't developed. It simply doesn't ship before less-perfect software captures the market.

[Dillweed]

An intresting analogy, a tad simplistic tho.

[Other Means]

Beware of arguing by analogy. That way madness lies.

[Peter Cairns]

However there is the Skyhawk analogy'

When the US had acompetition for a new carrier attack aircraft, everyone but one group submitted a prop aircraft.

However the winner which went on to be the A-4, was the only jet, a single engined delta that outperformed all the rest.

40 years later the last uprated A-4's could fly far further with a much greater payload at higher speed, but the US didn't buy them, why?

Because it went for the JSF, because despite the performance of the late A-4, it just had reached the end of it's development, and even with the latest technology, it would never be stealth or VSTOl.

Sometimes you have to go back to basics and design from the ground up.

Peter.