Merb and Rails are merging?

[Kyle Drake]

http://weblog.rubyonrails.com/2008/12/23/merb-gets-merged-into-rails-3

What's the skinny guys? Is this an early April Fools joke, or the real deal?

-Kyle

[Michael D. Ivey]

> http://weblog.rubyonrails.com/2008/12/23/merb-gets-merged-into-rails-3
>
> What's the skinny guys? Is this an early April Fools joke, or the
> real deal?

http://yehudakatz.com/2008/12/23/rails-and-merb-merge/

Real deal.

[Michael Klishin]

On 23.12.2008, at 23:25, Kyle Drake wrote:

> What's the skinny guys? Is this an early April Fools joke, or the
> real deal?

Real deal.

MK

[Zack Ham]

Coming from Rails, I really hope that merb's naming scheme wins out.
Application beats ApplicationController and before beats
before_filter. Also would love to see merb-action-args come along in
this merge. Crazy times ahead.

Zack

[Michael Klishin]

On 23.12.2008, at 23:46, Zack Ham wrote:

> I really hope that merb's naming scheme wins out.
> Application beats ApplicationController and before beats
> before_filter.

I personally think Rails convention will win simply because there is
more code to change otherwise. But it does not look like a problem to
me.
There are much more important things like making ActiveSupport monkey
patch less aggressive and perform better, make ORM finally pluggable,
etc.

MK

[Jon Hancock]

a few minutes ago I finished my rewrite of shellshadow in merb
1.0.6.1. This took on and off a few months as merb 1.0 was in flux.
Its late (6am in Shanghai), I'm very tired, I haven't finished my
Christmas shopping yet, but I checked the merb google group before
going to bed. I see this merger announcement.
ok, fine. I'll have to sleep on it and then I guess I'll wait a few
months to see what happens.

But what I went through:
2007: launched shellshadow in rails 1.2.3. I was very unhappy with
the experience (read below to see why I had some insight into
frameworks)
2008 - April: re-launched with a rewrite of shellshadow in merb 0.9.2
+ datamapper 0.2.5
2008 - December: re-launch shellshadow rewrite in merb 1.0 + dm 0.9.8

Do you know how tired I am of following the framework curve? Let me
explain...

I wrote the world's first ORM in smalltalk in 1988. From then through
the early 90s I wrote and re-wrote several frameworks until I evolved
a "full-stack".
In '94, I wrote the world's first full-stack app framework,
"patternWare", in Smalltalk.
I re-wrote patternWare in Java starting in '98 - 400,000 lines of Java
code, an ORB, the whole frickin' kitchen sink. I eschewed and blew by
the promise of J2EE. Yes, it was DRY, and all the other goodness the
ruby worlds thinks it invented. Well, ok, the smalltalk stuff was
much DRY-er than the Java ;)
The above frameworks were used by many Fortune 100 companies. Big
honkin' enterprise apps that required a mature easy to use framework
that could be taught to old-school COBOL programmers. F/OSS
principles didn't apply then. My customers didn't pay for beta code.
I literally spent a decade trying to explain to anyone that would
listen the importance of this thing I was calling an "application
framework".

That all started 20 years ago...Now I'm tired. I'm too old to hack
away on frameworks. There are plenty of smart young people to do
this. I may have some value to add to this community. Maybe not.
But really I'm just tired and all I want to do is write a webapp that
has sound underpinnings. Due to the permissive nature of the web,
this requires frameworks.

All I ask is for the community to take it easy. Do you really think
you can fix all the outstanding bugs in merb 1.0.x while putting such
an extraordinary effort into Rails 3? The answer is yes, given a long
enough timeline. But no, if you're too aggressive.

Please try to take it easy on people, like me, who's main interest
these days is publishing a webapp, not following the edge.
Frameworks, like math, is a young man's sport. I've had my days of
existential scheming in the snow.

thanks for your continued help and understanding. And thanks for
letting me have my little rant here ;)
My biggest fear about this merger is that I won't have such close
access to the people that make merb great. I feel the contact will be
drowned in a sea of Rails.

take care and Merry Christmas to you all.

- Jon

On Dec 24, 4:55 am, Michael Klishin <michael.s.klis...@gmail.com>
wrote:

[Zack Ham]

For better or worse Jon I think web development has always been a fast
moving target. It seems we are always targeting compliance with
evolving and unimplemented standards. We are continually pushing the
capabilities of browsers. We are taking a medium that many of us
remember as being basically documents with links, and turning it into
a thin-client application runtime environment. It is crazy and that
is not changing. I agree with you though, and hope that this merge is
handled with careful deliberation, as it could have a variety of
outcomes and I truly do hope for the best.

- Zack

[Ryan Bigg]

I also help merb wins. There's been a few instances where, in Rails
(and maybe in merb too) that people have had to change their code
(remember the "undefined method scaffold" saga?) and for a (major)
major release people should expect there to be some changes they will
have to do.

-----
Ryan Bigg
Freelancer
http://frozenplague.net

[cult hero]

Well... how happy I am about this is purely a matter of seeing how
things pan out. Merging with Rails = lots of developers and a bigger
user community. This generally equates to more and better
documentation, more general acceptance, more choices for hosting and
easier job opportunities. There is a reason PHP flourished and
continues to do so. Weee!

It also means extended bureaucracy and in general, it's harder for a
single vision (which is often a good thing) to do its thing. I've
really been impressed by Merb's core team and their goals and vision.
When I saw Yehunda was also a part of jQuery (something a discovered
about a year ago) I was excited because I loved jQuery's philosophy.
I'm worried that this is going to eventually get trampled under the
weight of the Rails community.

However, I also trust that the core team has good reasons for doing
this and wouldn't make the merge if it were to merely produce an
incrementally different version of Rails versus a radically different
version.

We'll see I guess. I suppose you could say I'm cautiously optimistic.

What I'm most curious about are the codebases. Is Merb itself becoming
the foundation of Rails 3.0 or are the Merb guys bringing the Merb
philosophy to the existing Rails codebase? I've seen forks become the
new standard or get merged back into the original project, but wasn't
ever a fork, was it?

If this is pulled off well, it'll be huge since it'll disarm one of
the primary attacks on Ruby and the web—performance and scalability.
I'll also be too happy to see all this hit with Ruby 1.9 (the unicode
improvements are HUGE for the work that I do). I'm really hoping the
next few years see a large increase in Ruby usage.

I actually figured something like this would happen. Either Merb would
eventually eclipse Rails or the two would merge. And, for those
worried about competition... there are always forks. This IS open
source, after all. If Rails 3.0 becomes a dog, another project will
rise to meet demands if there are a sufficient enough number of them.

I'll be happy as long as Merb continues to at least get bug fixes and
the time I spend with Merb (I'll still be using it AT LEAST until
Rails 3.0 is upon us) isn't wasted should I choose to make the
transition to Rails 3.0.

[Michael Klishin]

On 24.12.2008, at 3:28, cult hero wrote:

> What I'm most curious about are the codebases. Is Merb itself becoming
> the foundation of Rails 3.0 or are the Merb guys bringing the Merb
> philosophy to the existing Rails codebase

It makes sense to move changes from a younger project to the more
mature one.

I want to ask community members is to calm down a bit and believe that
newly formed rails-core-and-rails-more team can figure out a good way
to make Rails 3 shine. But roadmaps for merges like this one are not
done in one day.

MK

[Tze Yang Ng]

I was shocked when ibm sold its pc sector to lenovo a few years back
... and so am i abt merb's merging into rails-3 ... anyway, thinkpad
still rocks, i guess so will this merge. Merging isn't a bad thing.

:] TY

==

--
http://ngty77.blogspot.com

[SB]

I wasn't exactly thrilled by this announcement. To be honest if feels
like EngineYard is preparing for tough times by spinning off/cutting
back all the open source projects they've been sponsoring (thanks by
the way). While getting merb into rails is a good thing, merging the
two is going to be a long road of compromises filled with heartbreak
for people who saw merb as "rails done right". Rails, while wonderful
and pioneering, was designed and written by php and java refugees who
were learning the language as they went. Rails source code seems
convoluted and dare I say ugly when compared to merb and the
organization of files just confusing. Merb to me was rails with 20/20
hindsight.

Just looking at these two side by side in a browser gives me serious
pause.

http://github.com/wycats/merb/tree/master

http://github.com/rails/rails/tree/master


I'm sorry if I come off as a negative naysayer but it just feels like
an independent film director whose work I respected finally got the
recognition he/she deserved but as a result signed on to direct Rocky
8. At the application level I think the disruption will be minimal
for newer projects since we're talking about the same MVC philosophy
and everything being ruby.

I see big wins all around for rails but what's in it for the merb
community?

Bigger community?
Merb already had the cream of the crop and we all saw how badly the
signal to noise ratio degenerated as the influx of developers
increased. Merb was like the study hall of a big city public school
where the smart kids hung out and bettered their minds rather than
chase girls and popularity.

Better documentation?
Maybe but the lack of solid documentation for merb encouraged many to
look under the hood and discover clean and concise source code. Plus,
if you argued that "specs are the docs" merb is in some ways much
better documented than rails at the source level.

Wider industry adoption?
Yellowpages.com and a bunch of other established rails sites migrated
to merb and it's only a time before any best kept secret stays kept.
There was enough adoption of merb to sustain developer interest in
keeping the framework moving forward.

This all comes as a shock just as I thought things were settling down
with merb 1.0.x.

Sam

PS I use rails at work and merb at home. I don't have anything
personal against rails. Just sad to lose a choice. My bet is merb
will eventually resurrect itself sometime before or after Rails 3.0.

[Rich Morin]

At 03:55 +0300 12/24/08, Michael Klishin wrote:
> It makes sense to move changes from a younger project
> to the more mature one.

I don't have any specific knowledge about either of the
implementations, but I'd like to push back on this a bit,
based on general software experience.

My impression is that the Merb core is smaller, cleaner,
and better documented than the Rails code. Adding Merb
code to Rails could therefore leave islands of clean code
in a sea of mud.

In contrast, if the Merb is used as the starting point
and folks like Ezra and Yehuda get to vet any additions,
I could believe that the result would be quite sanitary.
That said, however, I defer to their judgement...

My other concern is one of attitudes. The Merb camp has
a refreshingly user-centered attitude (where the user in
this case is the Merb developer). Basically, they act as
if their goal is to make the developer's life as easy as
possible. I haven't seen this in the Rails approach, to
date, so I worry that things could change for the worse:

"If man could be crossed with the cat it would
improve man, but it would deteriorate the cat."

-- Mark Twain

-r
--
http://www.cfcl.com/rdm Rich Morin
http://www.cfcl.com/rdm/resume r...@cfcl.com
http://www.cfcl.com/rdm/weblog +1 650-873-7841

Technical editing and writing, programming, and web development

[Justin Reagor]

Bigger community?

Better documentation?

Wider industry adoption?

PS   I use rails at work and merb at home.  I don't have anything
personal against rails.  Just sad to lose a choice.  My bet is merb
will eventually resurrect itself sometime before or after Rails 3.0.

Your points are dead on in my book... with a dash of "WTF I smell fish?". Can you imagine what Merb would be on its own by the time Rails 3.0.x is perfected?

I also can't see other frameworks like SproutCore, adopting Rails. Unless Rails 2 is trashed and completely replaced by Merb. In which case, why does Rails get the title and tech just handed to it? Because we need the logo rights?

:: Justin Reagor

:: just...@gmail.com

[Jacques Crocker]

Couple points:

- Merb name doesn't make all that much sense anymore. Mongrel + Erb?
Anyone still use mongrel? Rails is a good name, despite the technology
baggage.

- Rails has amazing industry adoption for such a offbeat Ruby
framework. Books + IDEs + consulting are huge for a project that a
young Danish punk put together in his spare time. Pretty good reason
to

- I'm extremely excited about this news. Having a small niche
framework is fun for a while, but not so much when you have no one to
write patches+documentation, and constantly find bugs in stuff that
you really expected to work. Having a larger user base and ecosystem
is always better than not.

My only concern is compromises. I do believe rSpec will win out as I
think even test::unit users agree its a better organized framework.
I'm hoping that they handle Rails 2.x compatibility via a plugin, and
not keep supporting the dead weight features that Merb provides
superior approaches to.

My 5 biggest frustrations with Rails are:
- render / redirect spaghetti code. each controller action should
return a string
- concat and erb buffers. every view helper (including ones that take
blocks) should return a string
- _url routes. unified url generation is vastly superior. helpers
should use strings
- rails gem plugins. these need to be deleted and replaced by Merb's
superb new dependency management
- no good pluggable apps solution (embraced engines for many months.
it was a bad many months)

The reason I switched to Merb for all new project development is that
it addressed all these with really slick solutions.

As long as most of these issues get addressed in Rails 3.0 (which I
plan on wholeheartedly contributing to), I wont have any problems and
the additional benefits of the merge will pay off (bigger less
fragmented community, better documentation, more jobs, etc).

Really looking forward to the Roadmap. This is a big issue that will
impact the future of many Merb developers who are currently working on
projects/plugins/patches that now may be disposable. The more open the
transition, and the earlier the roadmap, the better.

Thanks!

> :: justi...@gmail.com

[Ahsan Ali]

Quite a shocker, and I have to agree with SB and Rich Morin.

I don't want a huge framework. I want a perfect parts that I can assemble for usage, and Merb supplies that. I want to keep it small.

I really don't see how such a complicated and bandaged piece of software like Rails can be improved by merb, unless its "Rewritten".

Somehow I get the feeling that this decision has been made for all the wrong reasons (commercial).

[Matthew Beale]

Ok, so I'm facing this with a smile on my face.

I didn't mind being pushed into merb at first. The benefits once I
worked around the smaller feature-set were significant enough, but we
sunk a lot of time into making merb 0.9 dance (porting in house plugins,
workling). The merb 1.0 transition was terrible though, and we had to
sink a whole bunch more time into not losing our upgrade path (you
*need* to use a thread and worker for deployment? gems changing without
new version numbers?).

And now a THIRD transition to something new this spring. I really hope
this one is a damn bit smoother. I guess it's the last though.

That's my smoldering outburst. Time to grin and bear and maybe remember
your Joel Spolsky:

http://www.joelonsoftware.com/articles/fog0000000069.html

I'm very excited to see the great parts of Merb merged into the Rails
codebase. I expect great things from great people working together on
great code!

--
Matthew Beale :: 607 227 0871
Resume & Portfolio @ http://madhatted.com

[Jon Hancock]

I share your hopes TY. But I have to disagree that the ThinkPad still
rocks. The X and T series are still very good and I recommend them to
many. But from a design perspective, they lost focus. Just look at
the keyboard difference between a T42 (IBM) and a T60 (Lenovo). Its a
clear case of every committee getting their favorite key added to the
keyboard. Same thing with the X series.
I'm a "less is more" type of guy.
Jon

[Ashley Moran]

On Dec 24, 2008, at 2:24 am, Rich Morin wrote:
> In contrast, if the Merb is used as the starting point
> and folks like Ezra and Yehuda get to vet any additions,
> I could believe that the result would be quite sanitary.
> That said, however, I defer to their judgement...

On Dec 24, 2008, at 3:03 am, Justin Reagor wrote:
> Unless Rails 2 is trashed and completely replaced by Merb. In which
> case, why does Rails get the title and tech just handed to it?
> Because we need the logo rights?

Rails 3 = Merb 2 + a transitional Rails compatibility layer, right?

At least I hope that is the intention, or there will very soon be
hundreds, if not thousands, of upset, disillusioned Merb developers
around.

Ashley

--
http://www.patchspace.co.uk/
http://aviewfromafar.net/

[Samo Korosec]

Hey, at least DHH ends his "adressing the concerned 4%" post at

http://www.loudthinking.com/posts/36-work-on-what-you-use-and-share-the-rest

classy;

"So kumbaja motherfuckers and merry christmas!"

First there's an annoucement that feels like a slap in the face to the
merb community and then we get a nice taste of what's coming, yay!

[Khaled al Habache]

Well, I was thinking the same way as Ahsan Ali, would we lose the lightweight framework?? But it seems the main effort will be to make smthing called "rails-core" , similar to merb-core....
Let's wait and c....

___________________________
Eng Khaled alHabache
http://www.khelll.com

[mikhailov]

What you think about new logo?
http://www.railsgeek.com/2008/12/23/rails-3-rails-and-merb-merge

On Dec 24, 4:21 pm, Samo Korosec <smoof...@gmail.com> wrote:
> Hey, at least DHH ends his "adressing the concerned 4%" post at
>

> http://www.loudthinking.com/posts/36-work-on-what-you-use-and-share-t...

[Thomas Sullivan]

New to this list - after finding merb when looking for a smaller
footprint than rails has. (I started, like many others, learning ruby
through rails. I know there's a good tool for every job, and I'm
sold, but here in the corporate environment, it's harder than ever.)

I'm not 100% sure I'm happy about this merge, however I do have to say
that the respect I have for rails has coupled with hearing about the
merge - has given me respect for MERB, too. (Have yet to tinker with
it, finding the time during Christmas is hard.)

I'm still going to learn myself the merb philosophy and see what
happens. I want that smaller "core" framework ideology.

[Rue the Ghetto]

While I do feel like this could be a good thing, I also feel like the rug has been pulled out from under me.

As the lead developer for a small Ruby web shop, I was able to push Merb into a couple interesting commercial projects. We have projects dating back to version 0.4.X. In the Spring, I built a green PR site for Whirlpool/Method that calculated the water quantity and usage of a user's household. It was so calculation and database heavy that it really shined under Merb. Sadly the project died after it was built for lack of marketing funding. 

Afterwards, we built greenwala.com. That is when the decision to use Merb turned into an investment in the framework. We had an excellent CMS built in Rails that met all our client's needs, but decided to rebuild in Merb because our company believed in Merb's speed, modularity and philosophy. We eschewed the numerous rails plugins that would have made our life so much easier, because we wanted to build a new set of tools for Merb. We struggled with openid, pagination, and facebook, using a much smaller pool of competent developers. We had plans to extract our CMS and possibly the social network into a slice. Just recently I started extracting the generic parts of facebooker into a gem to open source. That way there could be an adjunct  rails/merb/mack/whatever gem that worked with framework specific items. After the announcement, doing that kind of work seems foolish.

Although I haven't made significant contributions to datamapper or Merb, I have been an ardent Merb user and an evangelist for this new little framework that could. Perhaps I flatter myself to think that Merb owes people like me, early adopters, some notice or input before this kind of change. 

I have a few questions for the new core team that will help me figure out where to invest my coding and open source time:

1) What is the process going to be like for adapting existing Merb plugins to the hybrid framework?

2) Can slices be used in Rails 3? And if so wow will slices have to be changed to fit in?

3) Will there be a performance hit for using legacy rails plugins?

Thanks for listening ... really!

[Michael Klishin]

On 24.12.2008, at 21:52, Rue the Ghetto wrote:

> 1) What is the process going to be like for adapting existing Merb
> plugins to the hybrid framework?

Right now all the discussions involve keeping compatibility as much as
we can. It is sometimes good, sometimes bad. We will see how it goes
in the future.

> 2) Can slices be used in Rails 3? And if so wow will slices have to
> be changed to fit in?

Slices/engines/you name it will be one of the crown jewels for Rails
3.0. Most ideas I heard of so far come from what slices taught us
through Merb's evolution.

> 3) Will there be a performance hit for using legacy rails plugins?

It depends on how your plugin is written. No framework can speed up
slow code that comes from 3rd parties.

MK

[cult hero]

Is there really a point to asking specific questions about technical
details right now? I mean, has any code even been written yet or is
the exact plan that clear. It's like some new movie has been announced
and before the script is even written people are asking who is going
to be the lead actor.

I cannot stress enough, as people worry, Merb is open source. It's not
going anywhere. DataMapper is its own project. It's still there. If
Rails 3 doesn't satisfy me, for instance, I'll use Merb and put a
community together to keep it going. This is why open source and free
software is GOOD.

On Dec 24, 11:06 am, Michael Klishin <michael.s.klis...@gmail.com>
wrote:

[Matt Aimonetti]

@cult hero, we are still working on the technical details so we can't
really answer your questions. I'm working on a blog post which will
hopefully address the concerns I heard in the last 24 hours.

We believe most merbists will be really happy with rails 3.0 since
they will get everything they like in Merb + Rails 3.0. You will need
to trust us and keep us honest. We believe we can do it, and you
trusted us with merb, so trust us with Rails.

If Rails 3.0 won't have what made Merb so special then it will be a
total FAIL and I won't blame you if you want to go back to the merb
code base and work from there. However, with 11 smart people on the
project believing they can do it, I strongly think that the odds are
on our side.

- Matt

[Ryan Bigg]

>
> Is there really a point to asking specific questions about technical
> details right now?

No one knows. Things change. Wait for it to mature and then see what
comes out. I wouldn't be too worried about the technical details
anyway since there's a bunch of competent people from the merb camp
and a bunch of competent people from the rails camp working on the
same project together. Generally (in a logical world) the sum of two
positives is positive.

[Brian LeRoux]

Yeah adding more people should really speed things up and help
innovation. Makes sense right?

[Ryan Bigg]

More people to bounce ideas off, and more people to disagree with you.

[Matt Aimonetti]

I hear some sarcasm in this comment ;) However, adding more people
with different tasks will def. help a lot. We are bringing our
expertise and we will work on separate parts of the framework. I'm
positive that will help.

- Matt

[Ahsan Ali]

There are two aspects to this merge: Technical and Cultural.

1) Technical - the upgrade path for mebists; how much of 'rails 3' will be merb; what each project gains from the other; community support for current Merb users.

2) Cultural - 'good ruby code' as Matt tweeted (e.g less metaprogramming); less fandom and more geekiness (merb); good communication with framework users (as in merb); specifying a clear protocol for contribution.

- highandwild

[Yehuda Katz]

On Wed, Dec 24, 2008 at 10:52 AM, Rue the Ghetto <ru_g...@rubyghetto.com> wrote:

While I do feel like this could be a good thing, I also feel like the rug has been pulled out from under me.

As the lead developer for a small Ruby web shop, I was able to push Merb into a couple interesting commercial projects. We have projects dating back to version 0.4.X. In the Spring, I built a green PR site for Whirlpool/Method that calculated the water quantity and usage of a user's household. It was so calculation and database heavy that it really shined under Merb. Sadly the project died after it was built for lack of marketing funding. 

Afterwards, we built greenwala.com. That is when the decision to use Merb turned into an investment in the framework. We had an excellent CMS built in Rails that met all our client's needs, but decided to rebuild in Merb because our company believed in Merb's speed, modularity and philosophy. We eschewed the numerous rails plugins that would have made our life so much easier, because we wanted to build a new set of tools for Merb. We struggled with openid, pagination, and facebook, using a much smaller pool of competent developers. We had plans to extract our CMS and possibly the social network into a slice. Just recently I started extracting the generic parts of facebooker into a gem to open source. That way there could be an adjunct  rails/merb/mack/whatever gem that worked with framework specific items. After the announcement, doing that kind of work seems foolish.

Although I haven't made significant contributions to datamapper or Merb, I have been an ardent Merb user and an evangelist for this new little framework that could. Perhaps I flatter myself to think that Merb owes people like me, early adopters, some notice or input before this kind of change. 

I have a few questions for the new core team that will help me figure out where to invest my coding and open source time:

1) What is the process going to be like for adapting existing Merb plugins to the hybrid framework?

The current plans involve porting plugins like merb-helpers to Rails3, and providing transitional releases (and probably a transitional plugin for Rails3) to aid in the transition. How does that sound?

 

2) Can slices be used in Rails 3? And if so wow will slices have to be changed to fit in?

3) Will there be a performance hit for using legacy rails plugins?

Thanks for listening ... really!

On Wed, Dec 24, 2008 at 2:49 AM, mikhailov <mikhailo...@gmail.com> wrote:

What you think about new logo?
http://www.railsgeek.com/2008/12/23/rails-3-rails-and-merb-merge

On Dec 24, 4:21 pm, Samo Korosec <smoof...@gmail.com> wrote:
> Hey, at least DHH ends his "adressing the concerned 4%" post at
>

> http://www.loudthinking.com/posts/36-work-on-what-you-use-and-share-t...

>
> classy;
>
> "So kumbaja motherfuckers and merry christmas!"
>
> First there's an annoucement that feels like a slap in the face to the
> merb community and then we get a nice taste of what's coming, yay!

--
Yehuda Katz
Developer | Engine Yard
(ph) 718.877.1325

[Matt Aimonetti]

Some of you still seem to have issues with the merge so I decided to address the most common concerns:

http://merbist.com/2008/12/25/merb-rails-merge-or-why-should-merbists-be-happy/

Let me know if you have any questions.

 - Matt