Choose Your Rut Carefully

Once seen on a sign beside a major dirt road in the Australian outback:

"Choose your rut carefully. You will be in it for the next 200 miles."

I have often used ChooseYourRutCarefully in advocating that large (and small) companies should standardise on Java for all new serious software development - as the least of all evils in the software rut wars. See http://www.objective.co.uk/events/presentation/9803_silverstream_city/sld015.htm (BrokenLink) and the slides following it for this line of reasoning in early 1998 (part of Objective's collective slideware but given very effectively by MikeStorey that day).

Then I gave up my strategic consultant pretensions and went back to being JustaProgrammer (and WikiPoster?) using Java plus a smattering of two cool untyped scripting languages called Python and Vision.

I still think that in the light of TheMostWidelyUsedProgrammingLanguageAtAnyLevel and the GreatestMarketingCompanyInTheWorld it is very important to be realistic about the limitations of our choices at the start of this brave new millennium. But I'm also having second (or maybe third) thoughts now about my advice. See the first trickle of this in NewLanguagesForXp.

Anyone got time to reflect in (or on) their current rut?

--RichardDrake

See StandardsEnforcement for a more cautious view on corporate standardisation than the one above.

Ruts are fine and good, but eventually you'll head into the ocean if you keep going to long.

I'll be surprised if there are many ruts leading out into the ocean.

Following a rut means repeating other traveler's decisions. With 70% of projects failing in the industry, I think I'm better off breaking a new path. -- JeffGrigg

You (and I) may well be better off. But what happens to the overall 70% failure rate when you filter the projects with "where number(previous projects where this language and tools have been used) = 0"? [FirstOoProjectDisasters] I have a strong hunch that it goes towards 100% (in fact I seem to remember EdYourdon and others have some stats to prove it).

A big component of "the rut" above refers of course to the maintenance of operational systems in whatever language they were originally written. As all those Y2K audits proved, "successful" systems (ones that at least are used on an ongoing basis) often last a lot longer, maybe decades longer, than everyone was kinda expecting. You (and I) don't get to break a new path so easily here.

--RichardDrake