The leaden weight of the legacy user
Ten years ago in my earlier life as a tech journalist I remember visiting Microsoft Research and being told that they were investigating all kinds of alternative ways of visualising and navigating through your computer data, harnessing the increasing power of desktop computers. Why is it then that even when the next major Microsoft OS revision emerges we will still face the "venerable shell metaphor of a desktop with overlapping windows"? And why, more to the point - the original impetus to my post - does SPSS 11.0 for Mac OS X still limit variable names to eight characters??!! I mean when Apple launched the Mac more than 20 years ago it already supported file names longer than 8 characters. Even PCs allowed you a three character extension to eight chars!
Update: … and I just discovered you can’t start the variable with a number (like the question number, say).
I know the answer why of course - backwards compatibility - I didn’t spend a year doing the LSE’s MSc in New Media, Information and Society for nothing… Nonetheless… Aargh!
September 15th, 2005 at 4:16 pm
Imagine the land of IBM mainframes. Restrictions that made sense 40 years ago when OS360 was launched are still there embedded in the code.
But the PC problem is several million times larger (there being more PCs and no easy way to convert data).
I remember to study deforestation and global warming they had to go back and look at the early 1970s Landsat photos, the first photos of the whole world with that data.
However they had to rebuild the machine to read the data from scratch– there was no other way to do it.
The phone system is another massive legacy problem: my parents still had dial phones until a few years ago. And of course we are tied into the basic AT command sets etc. that AT&T established in the 1930s…
‘path dependence’ as economists call it: where you finish up is conditioned to a large extent by where you happen to start (QWERTY keyboards being the canonical example).