What's the current standing of Rant vs. Rake?

[Chris Pearl]

Rake seems to be used widely and extensively. Rails uses it pretty
extensively, as do quite a few of the tools I'm currently reviewing
for the purposes of another thread [1].

So what about Rant [2]? It's certainly not dead: latest release was on
November 17, 2006. And it is claimed to be more featureful than Rake
[3].

Where is Rant being used (if at all)? Why does Rake seem to be the
overwhelming choice?

-Chris

 

[Chris Pearl]

Further Investigation clears up some of the mystery.

The following ML post [1] claims that Rant
covers some major use-cases which Rake doesn't. Key quote: "Rake
appears to have no concern with anything outside the Ruby community".
In other words, Rake is dedicated to building Ruby software system,
and not interested in building anything else, while Rant is.

The author of that post was familiar with Rake [2], and later became a
lead developer of Rant. So I assume he knows what he's talking about.

That said, the above provides a solid conceptual ground for the
overwhelming preference of Rake over Rant in both Rails and the static
site generation systems I'm currently reviewing, all of which are
focused on very simple and/or Ruby-specific tasks.

As for more practical grounds, this ML post [3] reveals that as recent
as six months ago, Rant was experiencing serious existential threats.
The sole developer announced he ran out of time and was not planning
any further development. As Rant was at that point (and still is) in
beta state, stagnation meant death. And indeed the major discussion
channel of the Rant community was fizzling out.

Under such conditions, it is no wonder that Rant would be perceived as
too risky a choice for anyone but those who actually needed the
non-Ruby-related features it offered over the stable, mature and vital
Rake. At least three of these later sort joined the Rant development
team soon after the quoted discussion took place. Since then, Rant has
been making pretty steady progress. I'm not sure it would become a
tempting alternative for those who can do with Rake, but it certainly
looks poised to becoming more appealing for those who need the extra
features and/or are looking for a build tool for non-Ruby-related
systems.

-Chris

[1] http://rubyforge.org/pipermail/make-cafe/2006-August/000089.html

Most of the information relevant to our discussion is summarised in
the the second-to-last paragraph:

"So I have arrived at Rant. Based on Ruby so easy to work with. Almost
but not quite a superset of Rake. Interested in building things other
than Ruby software systems. Seems like a better choice than Rake. But
is it? This is a question I have to ask myself and have a realistic
answer; I need to feel comfortable about investing the effort to move
from Rake to Rant. Apart from the fact that Rake has an article by
Martin Fowler and is deeply internalized with the Ruby community, it is
stable. It doesn't seem to have any community but it does do the job
and lots of people are downloading it. Rant appears to be technically
superior but it has no community and no publicity machine. The analogy
here is between VHS and Betamax -- VHS was inferior but won out because
of the publicity and expectation management. Currently Rake is winning
over Rant because it appears as a download on the Ruby front page."

[2] http://rubyforge.org/pipermail/make-cafe/2006-August/000083.html
[3] http://rubyforge.org/pipermail/make-cafe/2006-August/000086.html

 

[Daniel Berger]

Further Investigation clears up some of the mystery.

The following ML post [1] claims that Rant
covers some major use-cases which Rake doesn't. Key quote: "Rake
appears to have no concern with anything outside the Ruby community".
In other words, Rake is dedicated to building Ruby software system,
and not interested in building anything else, while Rant is.

The author of that post was familiar with Rake [2], and later became a
lead developer of Rant. So I assume he knows what he's talking about.

That's interesting, considering the Rake tutorial
(http://docs.rubyrake.org/read/book/1) actually covers building C code.
Does the OP mention *which* use cases Rake doesn't cover? I'd be
curious to know.

That said, the above provides a solid conceptual ground for the
overwhelming preference of Rake over Rant in both Rails and the static
site generation systems I'm currently reviewing, all of which are
focused on very simple and/or Ruby-specific tasks.

I think it stands to reason that most people would use Rake for Ruby (or
C). Just as most people use Ant for Java even though, in theory, you
could use it for something else. Such is the nature of things.

As for more practical grounds, this ML post [3] reveals that as recent
as six months ago, Rant was experiencing serious existential threats.
The sole developer announced he ran out of time and was not planning
any further development. As Rant was at that point (and still is) in
beta state, stagnation meant death. And indeed the major discussion
channel of the Rant community was fizzling out.

That would be a shame. I like the idea that there's an alternative out
there if we need it.

Under such conditions, it is no wonder that Rant would be perceived as
too risky a choice for anyone but those who actually needed the
non-Ruby-related features it offered over the stable, mature and vital
Rake.

I'm not really sure what risk there is in practical terms, given that
gems are the primary way most people get their code these days.
Gem#add_dependency('rant').

I looked over the Rant project a bit, btw. Superficially it looked very
similar to Rake, and I didn't see a compelling reason to switch from
Rake. That doesn't mean one doesn't exist, however. Rake is certainly
not perfect IMHO, but it's "good enough".

At least three of these later sort joined the Rant development
team soon after the quoted discussion took place. Since then, Rant has
been making pretty steady progress. I'm not sure it would become a
tempting alternative for those who can do with Rake, but it certainly
looks poised to becoming more appealing for those who need the extra
features and/or are looking for a build tool for non-Ruby-related
systems.

-Chris

If there's a compelling reason to switch, then people will switch. If
not, they won't.

Regards,

Dan

 

[Chris Pearl]

Does the OP mention *which* use cases Rake doesn't cover? I'd be
curious to know.

In another thread, the author (Dr. Russel Winder) mentions the
building of LaTeX books:

"SCons supports this with builders that come as
standard with the distribution. Rake appears to have no ready-made
builders of any sort so I ended up writing my own functions to do the
job. Rant appears also not to have any ready-made support for LaTeX but
appears to have a more sophisticated view of things than Rake."

http://rubyforge.org/pipermail/make-cafe/2006-August/000083.html

-Chris

 

[stu]

Rake seems to be used widely and extensively. Rails uses it pretty
extensively, as do quite a few of the tools I'm currently reviewing
for the purposes of another thread [1].

So what about Rant [2]? It's certainly not dead: latest release was on
November 17, 2006. And it is claimed to be more featureful than Rake
[3].

Where is Rant being used (if at all)? Why does Rake seem to be the
overwhelming choice?

I use rant all time time for building C apps and such and have ran into
only 1 issue (its on the rant bugtracker), other than that its rock
solid. I dont use rake so cant comment over it. Rant does all I need it
to do. I dont consider it not production quality despite its version
numbering.

-stu

 

[Bil Kleb]

If there's a compelling reason to switch, then people will switch. If
not, they won't.

The ol' Beckhard-Gleicher model of change:

change if dissatisfaction * appeal * plan > cost_of_change

See also Moore's /Crossing the Chasm/.

Later,

 

[David Vallner]

--------------enig635D108236ED81085D937F81
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable

Seems like a waste of
time to create two very similar projects just because one doesn't have = a
feature or two.
=20

Yes, because it is unheard of that open-source projects would stay
separated or that one would fork because of a few features.

end # sarcasm

The projects they don't merge because there aren't compelling reasons
for the core developers of both of the projects to do so. (Avoiding
duplicated effort isn't such a reason if they're both hobby projects -
making the process of achieving goals as much of a motivation as finally
reaching them.)

David Vallner


--------------enig635D108236ED81085D937F81
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: OpenPGP digital signature
Content-Disposition: attachment; filename="signature.asc"

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (MingW32)

iD8DBQFFdyfqy6MhrS8astoRAq6VAJ42N5L63JbMmtgw3TBx9yVCWTYdowCeJsWC
jLnNp2H/gssp596MWfuyyqE=
=zGa4
-----END PGP SIGNATURE-----

--------------enig635D108236ED81085D937F81--