Monday, December 06, 2004
on the transformation of the software profession, etc.
update 2005.04.08: comments disabled. this particular post was starting to get spammed.
sort of like an “al gore disease” (his (attributed) “i invented the internet” comment). then again, i don't want to be casting al as a “big guy” ;-)
two sentences, and i'm off topic already ;-) what else is new. what i really meant to talk about was an article series i found at msdn while searching for "aspect oriented" patterns refactoring "visual studio" (that's a good search, try it, and variants without the visual studio part). i was looking for aop tools for visual studio / windows platforms, in the context of pattern / refactoring implementation.
what i found ? stuff on “software factories”. anyone who has worked with me in the last five years or so has heard me talk about how there needs to be a fundamental paradigm shift in the software (and hardware, for that matter) industry that begins to treat the discipline as an engineering undertaking, rather than a craft, and all that this implies to practitioners, organizations, end users, etc. it's not a palatable subject to some folks, and some have even argued that software development can never be engineering. don't let the “software engineer” job titles fool you - almost all the time that's just smoke.
so, like, here i find it, someone else picking up on the future, while i just sat on it. then again, i've always been a lousy salesman, no one ever seems to listen, and i end up being either a crank, or too far ahead of my time, or whatever.
anyway, the articles:
good stuff. i should have - would have - written on this subject long ago, but since i've been on sabbatical (;-)), the topic hasn't really come up much. and before that, i was blog-free. (yes, all this crap you see here, i used to annoy people with verbally ;-))
to end, just a mention that i suspect that aspects might be the1 method of direct implementation of patterns (rather than the original descriptive approach, or templates, or whatever). with automatic aspect and refactoring tools appearing for general use on many platforms, programming is on (yet) another threshold. it seems that those msdn articles mention an improved application of the aop concept, based on something closer to runtime weaving, but i still have to read them thoroughly to find out if this is an accurate interpretation.
not that all this aspect & refactoring & other more or less orthogonal analysis concerns is really the way we should go. that's all kind of top-down stuff, and seems to be kind of counter to agile software methodologies like xp. at the other (apparent) end of the spectrum is the bottom-up self organizing stuff represented by mass agent interaction. but i suppose, putting these things together, we just end up with more flavors and richer interactions of agents. we have to give up such fine grained control, and be content with creating the soup from which the desired functionality will emerge. such work in guided software evolution has already been proceeding for quite a while.
1 this is really pretty funny: i am so used to not using caps, and so oriented towards css, that instead of using the shift key and typing “the” in caps, i did it in lower case then surrounded it by a span with the text-transform:uppercase style attribute ;-) i left it that way, for posterity.
book crossing use as book loan tracker
no, not the one in the refrigerator ;-)
heyyyyy... i can use this (book crossing) to track books i lend out, not simply limited to releases to the wild. wish i would have thought of that before handing a dozen or so to my brother's family over labor day weekend. how about it, john, are those still hanging around ?
use #2: for used books sold (to bookseller, ebay, whatever) as well.
(bad?) idea: we could use something like this to track everything. which is kinda like where rfid is heading.