|
Authored by: Anonymous on Monday, December 10 2012 @ 07:48 PM EST |
Oh, I see. I still don't find that convincing.
In most cases (and they way I work) a programmer doesn't
execute the code in their head, instead they hold some
representation in their head. Inputs can be megabytes and
megabytes; there can be millions of steps. Clearly the
programmer doesn't go through all those from beginning to
end (that would be silly).
Similarly, and Graham doesn't mention it, programmers make
mistakes, pointing to this lack of execution. Bugs result.
Commonly any piece of software has lots and lots of bugs.
(Graham is romanticizing)
If the freedom to follow the computation by hand is a
problem, patent law could be changed to allow software
patents with the proviso that slow computations, on the
order of 1 step per second, are excluded. This would address
the concern with regards to execution-of-algorithms-in-your-
head (or with pen and paper). But it wouldn't address the
practical problem of patents, which is (1) patents are
useless for spreading knowledge with respect to software,
(2) there is too much variety for meaningful searches, (3)
it is difficult to draw lines, distinguish between software
functions. This is why I don't think this is a convincing
argument, the problem is of a different nature.
Also bear in mind, that nowadays, most people (my
impression) wouldn't even perform simple computations like
adding small numbers such as 17+68 in their head or with
pen and paper. Instead they use a calculator, or (shudder)
open a spreadsheet. It will be difficult to explain to them
the loss of freedom you are concerned about.[ Reply to This | Parent | # ]
|
|
|
|
|