Re: Differences between Org-Mode and Hyperbole

[John Wiegley <jwiegley@gmail.com>]

>>>>> Richard Stallman <rms@gnu.org> writes:

> The right way to integrate Org mode into Emacs would be to pry out each of
> those subfeatures and integrate it individually -- so that a user could use
> each of them either with or without Org mode. It is not too late for people
> to do this sort of thing, but it should have been done before.

However, the super-tight coupling of Org-mode's features allows them to
cooperate in ways that are usually "to the point", and sometimes elegant.

I find that when we try to develop separate submodules from scratch, and then
combine them, so many compromises get made during the effort to integrate
these modules (because who can design such a complex system right the first
time, understanding both the best way to abstract features, and the proper way
to expose their functionality?), that we end up with a huge, unwieldly system
that barely delivers on its promises. Some examples do come to mind...

I'm not saying Org-mode represents an ideal design. I'm saying that in the
real world, it represent an effective strategy for making a system that is
maximally *useful*. Hence it's extreme popularity. Even Carsten will admit
he's not a software architect by trade; he did what he did based on what he
wanted to achieve with Org-mode, and not based on engineering decisions.

Now that all the useful work has been done, and experiences gained, it could
be a good time to sift out some of the best of its functionality into separate
modules. However, I disagree with the assessment that it "went astray" by not
striving from separation from the beginning. I would even argue that some
projects that begin that way go astray by doing so.

-- 
John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2