Re: Top 10 for old farts

["Joseph B. Gurman" <gurman(at)ari.net>]

In article <MPG.13eb4d1b4b64e350989b77@news.frii.com>, 
davidf@dfanning.com (David Fanning) wrote:

[snip]

> In fact, I haven't written a program for a client in the
> past year that hasn't included at least one object,
> and sometimes it's easier to write the whole thing as
> an object. PSCONFIG, a program on my web page that is 
> widely downloaded and praised by people who use it, is
> an object program, although I would guess 95% of the 
> people who use it don't know that. Certainly most of
> the programs I add to my library in the future will
> be object programs, although I'm not fool enough to
> tell anyone that, since just the word itself is enough
> to send shivers up the backs of many IDL users. :-)

    I believe, I believe, I believe. I just have no need (yet) to write 
object code myself.

> 
> I've frankly pretty much given up the idea of writing
> an object book because (1) it is so damn hard to write
> a book, and (2) after going to all that trouble I thought
> only about a dozen people would buy it. (And I will hear
> from all 12 today, probably, pleading with me to reconsider,
> so desperate is the need for decent documentation.)

    Well, you might sell a 13th, just because you've piqued my interest, 
but I would put it in the "programming to learn" pile, right under the 
10%-finished Perl book, and the 0% finished Palm OS book. Frankly, I 
think I'd rather learn to write Palm OS apps (so I can analyze data 
anywhere), but that's another story.

> 
> The lack of good instruction is probably what is holding
> the adoption of objects back. I know you say you don't
> need them. The people in my course last week were adamant
> that they didn't need to know any widget programming, too.
> But in the end all they cared about was widget programming
> and increased interactivity with their data. I submit that
> objects have the same ability to transform how we work
> with our data.

    Careful what you say, Dave, because the truth can hurt. I _use_ 
widget apps every day --- that I specified to a good widget programmer 
years ago, and that get maintained sporadically. If you or your book can 
convince me I need some capabilities of IDL objects I haven't seen yet, 
I might ask a widget programmer to gin something up for me.... but it 
would have to be an awfully big something. You see, I went to the 
trouble (during a six-month sabbatical in France 8 years ago) to learn 
the original widget API.... and then they went and changed it. (In the 
meantime, I'd written some pretty neat widget code, IMHO.) In the words 
of Boris Badanov, "Fooey and double fooey." Just not enough incremental 
functionality to make me learn the whole, new interface when we had 
realtime data acquisition and reformatting software (the latter in IDL, 
no widgets) to write, prototype, and implement, and a satellite to 
launch.

> 
> Yes, there is more up-front cost. And I'm completely
> sympathetic with those scientists who feel they don't
> have time to figure it out. If objects are unfamiliar
> to you, it is just about impossible to learn about them
> from the material that RSI gives you. But in the end
> they do more for your science than what you are using
> now. That must be the bottom line. Someday, inevitably,
> you are going to be working with objects. Maybe you
> won't write them. But you sure as heck don't want to
> pay extra for them, either.

    Yes, I do want to pay "extra" for them, if for the same delta I can 
get QuickTime movies and interlaced GIFs instead. Honestly.

    As for the argument that RSI shouldn't charge on a 
by-not-so-useful-to-everyone feature basis, I view IDL rather like a 
power tool. If Black and Decker or Makita wants to add lots of bells and 
whistles to their new model of electric screwdriver, that's great, but 
I'd rather they simply make them all options that I can buy if I want or 
need them --- and by and large, they do. Now I know some people worship 
their power tools, but I find them fun and useful --- just like the 
parts of IDL I use. Well, maybe "fun" is too strong a word for some IDL 
"features," but....

    Since I leave for the IAU meeting in the UK and then vacation (and 
hopefully Intenret abstinence) for the next three weeks starting 
Wednesday, I won't dribble any more gasoline on this particular fire. 
The RSI lurkers can gauge how many people feel this way or that.

    Best,

                  Joe

P.S. My favorite IDL 5.3 "feature" (tested only in OpenVMS 7.2-1 so 
far): including the statement

device, cursor_standard = 68

in your IDL startup file reduces your available pixmap memory (e.g. for 
XINTERANIMATE movies) by 50%. What a hoot! Those programmers at RSI have 
such a sense of humor. (No such ill effect is seen in earlier versions 
of IDL.)