emacs-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ELPA] New package: GNU Hyperbole 6.0


From: Robert Weiner
Subject: Re: [ELPA] New package: GNU Hyperbole 6.0
Date: Sat, 16 Jul 2016 11:46:57 -0400

On Sat, Jul 16, 2016 at 11:11 AM, Stefan Monnier
<address@hidden> wrote:
> Elisp packages are typically fairly small, so complex structures are
> usually overkill.  Bigger packages generally benefit from being split
> into various smaller ones.

I see two sides to this.  If you create a large package with multiple
subsystems all connected through a single user interface, you could
split it into a master package with separate packages for the
subsystems upon which the master depends.  This increases modularity
and potentially allows application of the subsystems with different
UIs.  But the costs are fairly high: increased version management
complexity, separation of the UI from the subsystems, more work for
the maintainers and increased documentation difficulty.  Given that
virtually all Emacs work is on a volunteer basis, keeping the author
and maintainer complexity at reasonable levels to encourage further
work should be a significant concern.

XEmacs separated out a lot of packages from its core but then offered
a sumo tarball with everything included as one option.  I bet the sumo
option was a very popular download because users want easy installs
and a good set of features each time they commit time to work with a
piece of software and will typically choose this over modularity.  So
if users like integration and it reduces author complexity, that is
one reason to favor it.  Internally, things can still be modular and
then programmers can utilize the modules as they like.

Just my thoughts.

Bob



reply via email to

[Prev in Thread] Current Thread [Next in Thread]