Re: from ruby/RoR to Java (framework unknown) :(
Ryan Davis
On Jul 4, 2008, at 10:31 , S2 wrote:
> My company today decided to ditch ruby development and to develop
> new web
> applications only with Java or .net.
> Current RoR applications will be migrated to Java.
> Sigh. No more ruby for me (in office hours).
most people at this stage would be posting their resume, not a
question about "good" java frameworks. :P
phlip
> My company today decided to ditch ruby development and to develop new web
> applications only with Java or .net.
Quit.
Were you using RoR with unit tests?
> Current RoR applications will be migrated to Java.
Was there a rationale? Did an executive read a magazine article or something?
--
Phlip
Robert Dober
Maybe, but as someone who certainly is not like most people (with the
exception that I guess that many people are exactly like this, but
this would lead to a tangent I am afraid) I see an alternative.
Would you think you could some jar files like jruby.jar ? ;)
Cheers
Robert
--
http://ruby-smalltalk.blogspot.com/
---
AALST (n.) One who changes his name to be further to the front
D.Adams; The Meaning of LIFF
Robert Dober
> phlip wrote:
>
>> Quit.
>
> i probably will.
>
>> Were you using RoR with unit tests?
>
> ajax stuff that excited our clients and developers where quite happy.
>
>>> Current RoR applications will be migrated to Java.
>>
>> Was there a rationale?
>
Of course there was, you stated it yourself above, "developers were
quite happy", sorry for my pessimistic look at this.
Apart of the idea that jruby might come in handy this decision really
looks odd to me.
Maybe somebody wants to sink the ship? Well at least, if they do, they
seem to be good at it sigh.
R.
phlip
>> Were you using RoR with unit tests?
>
> no. everything was fine actually. apps where developed quickly, used "cool"
> ajax stuff that excited our clients and developers where quite happy.
You would have been faster and more ... assertive ... with unit tests.
Use assert_javascript to test Ajax.
Your Java and .NET offerings will limp far behind RoR's test abilities.
>>> Current RoR applications will be migrated to Java.
>> Was there a rationale?
>
> no.
You have a boss who is allowed to decree things without giving a rationale?
> probably. i think the main reason is that management does not want to have 3
> separate development lines (RoR, Java and .NET) for Web apps. I think they
> want to unify the development Frameworks/languages so they can relocate
> Staff from one project to an other without loosing to much "learn" time.
So, no pair programming, and they want to make knowledge transfer easier by
reducing the amount of knowledge required. Gotcha!
phlip
> Yes. I asked to get some devs and PMs together and get an analysis done on
> the migration, what we lose and what we would gain by dropping RoR for
> Java, but "for now we don't have the time. New development should be done
> in Java".
At my day job we use pure XP with RoR. We use promiscuous pair programming,
meaning we switch pairs every 2 hours. We use pure Test Driven Development, with
refactoring & continuous integration. We have a very high velocity and a
vanishing bug rate.
Since I got there, our team has increased to 6, and we will get 2 more as soon
as we remodel. I really doubt we could talk our executives into switching back
to Java, or switching from XP to something else... like "Agile".
Your worksite has issues that prevent your executives from seeing which systems
are working better than others. There's an old adage, "If I had 6 hours to cut
down a tree, I would spend the first 3 hours sharpening my ax".
Your bosses think the more strokes required to cut down a tree, the more
"progress" you made. That's what a 600 line configuration file is - 600 more
strokes with a dull ax.
>> So, no pair programming, and they want to make knowledge transfer easier
>> by reducing the amount of knowledge required. Gotcha!
>
> Exactly. I'm not sure I understand your point?
Exactly... you are not sure? (-;
If you are not pair programming, then knowledge is indeed hard to transfer at
team swap time. So your execs think they can reduce the total amount of
knowledge required. Yet more strokes to cut down that tree!
--
Phlip
Robert Dober
> Robert Dober wrote:
>
>> Of course there was, you stated it yourself above, "developers were
>> quite happy", sorry for my pessimistic look at this.
>
>
>> Apart of the idea that jruby might come in handy this decision really
>> looks odd to me.
>
> some other job. I think we/they are going to use the Seam Framework, but
> it's still not sure. They asked me to join the team who gets to evaluate
> the various available frameworks, but every time I look at a hibernate
> config file I get goose-flesh.
>
>> Maybe somebody wants to sink the ship?
>
>
>> Well at least, if they do, they
>> seem to be good at it sigh.
>
Sure, if you trust that they are honest, first thing to do is to ask
for their rationale.
Maybe, after all, it makes sense.
If they cannot share their agenda with you, I was probably a little
bit right, but if they can and do you might be able to
trigger a discussion.
If e.g. you were guessing correctly, and I guess knowing the ship you
are sailing on there are some chances that you have, that they are
trying to get their development sharing spectrum smaller, you might be
able to give them different options (jruby might come into play
again). You might want to figure out some market trends, development
times etc.
In an ideal world you might get a chance to present Ruby's strengths :).
But in order to reason with somebody at first you have to trigger a
dialog, do you think you are in a position to do that?
Well good luck from the bottom of my heart.
Cheers
Robert
James Britt
> No. jruby will not be an option. If that was the case I would not look for
> some other job. I think we/they are going to use the Seam Framework, but
> it's still not sure. They asked me to join the team who gets to evaluate
> the various available frameworks, but every time I look at a hibernate
> config file I get goose-flesh.
You need to be able to explain to them just why you get "goose-flesh".
No doubt everyone on this list is automatically sympathetic to your
plight, but truth is that much of what we see as obvious about Ruby is
not so for others.
I don't see how anyone can expect people to just be instantly
enlightened about the value of using Ruby; folks need facts.*
If Ruby (with Rails or whatever) is the objectively better choice then
you should be able to demonstrate that. There has to be something
tangible, something other than "I like it."
Get some stats, some code, some use cases, and build a solid case. The
best argument tends to center on money. Something that argues for
higher employee retention rate, or lowered training costs. If you can
show that using Ruby means faster turnaround with fewer developers and
lower maintenance, that in the end Ruby == more profit than Java or
Net, you should be OK.*
If you can't make that case, then why would you expect a business to
choose Ruby?
* Or not. Some folks will have ulterior motives for choosing one tool
over another, but you should start be giving people the benefit of the
doubt.
--
James Britt
www.happycamperstudios.com - Wicked Cool Coding
www.jamesbritt.com - Playing with Better Toys
www.ruby-doc.org - Ruby Help & Documentation
www.rubystuff.com - The Ruby Store for Ruby Stuff
Jim Morris
> My company today decided to ditch ruby development and to develop new web
> applications only with Java or .net.
> similar to RoR? Maybe something that emulates Activerecord and is as
> flexible, powerful and convenient to use as the ActionPack?
>
>
Rather than berate your management for the decision I'll try to answer your actual question ;)
I switch between RoR and Java regularly for different clients.
For Java I use the Spring framework almost exclusively. I do not use Hibernate as I prefer to use
springs DAO framework, and the various jdbc wrappers they provide, I like to get my hands dirty with
SQL, and you can make things very efficient that way too.
Spring has a fairly steep learning curve but is well documented and supported and has many
components, from JMS, WEBMVC, JDBC etc. you can pretty much do everything with the Spring framework.
I use a standard architecture where you have web servers, talking to middleware servers via JMS,
this separates all database access out of the web layer. Then the middleware layer is composed of a
JMS layer calling POJO's and the POJO's call the DAO layer which talks to the database, so
everything is cleanly separated. All database access is in the DAO layer, all Business logic is in
the POJOs. This makes testing pretty easy, I write JUnit tests (using hamcrest matchers) for the
DAOs, and for the POJOS, so I can test database independent of business logic and vice versa.
It is not as much fun as RoR, but it sure is much easier to scale for huge loads.
I use Eclipse as my IDE for Java development.
I would guess it takes me twice as long to write a given application in Java as it would in RoR, not
including the scaling issues.
I write all my support scripts and database management code in Ruby and use the Sequel GEM, so I
still get my ruby fix even when on a Java project.
Good luck
--
Jim Morris, http://blog.wolfman.com
Bill Kelly
From: "M. Edward (Ed) Borasky" <zn...@cesmail.net>
>
>> * Or not. Some folks will have ulterior motives for choosing one tool
>> over another, but you should start be giving people the benefit of the
>> doubt.
>
> accepted environments and projects.
Wait... have you ever actually endured much Java programming? <grin>
HaHaOnlySerious,
Bill
Eleanor McHugh
> Perl and Java on some simple linear algebra, and Java was the
> fastest. This was *long* ago -- before "Java 2"! So I wrote it in
> Java. The damn thing is still sitting in a CM repository somewhere.
> Nobody used it, because Java wasn't on the "approved list". If I had
> written it in C, I would have been a hero. Ah, well. :)
>
> That's the only Java code I ever wrote, but the language still
> appeals to me as a language. What I haven't endured is all of the
> frameworks, APIs, XML, etc. that have "evolved" around Java.
It started out a pretty sweet language: simple to learn and less
fiddly that C++. Unfortunately it's become progressively less friendly
with each successive release, not least because of the proliferation
of standard libraries. If Sun had started out with something a bit
more like Groovy I might even have remained a fan :)
Ellie
Eleanor McHugh
Games With Brains
http://slides.games-with-brains.net
----
raise ArgumentError unless @reality.responds_to? :reason
Robert Dober
Or Smalltalk, IIRC it was just of these *stupid* license issues that
they did not continue with Oak.
What a shame :(
R.
Aníbal Rojas
- Grails: http://grails.org/
- Seam: http://seamframework.org/
Happily since we left Java a few years ago we haven't had to look
back. I tried Grails and Seam at their early stages and is not the fun
Ruby provides but they do the work in a much better way than Struts
for example. They have very different approaches but they both follow
DRY and COC.
Good Luck,
--
Aníbal Rojas
http://hasmanydevelopers.com
http://rubycorner.com
http://anibal.rojas.com.ve
Eleanor McHugh
> Aníbal Rojas wrote:
>
>> - Grails: http://grails.org/
>
> I am being naive maybe, but I don't get the point of Grails. It's a
> different programming language and it tries to mimic rails. So the
> only
> thing it has of Java it's the JVM. So they had to completely rewrite
> rails
> in Groovy. Why not use the real thing at this point? Or jruby?
Because they like Groovy :)
It's a good language if the JVM is your only target, and it's still
close enough to Java that there's not too much pain working with both.
phlip
> Take a look at:
>
> - Grails: http://grails.org/
> - Seam: http://seamframework.org/
>
> Happily since we left Java a few years ago we haven't had to look
> back. I tried Grails and Seam at their early stages and is not the fun
> Ruby provides but they do the work in a much better way than Struts
> for example. They have very different approaches but they both follow
> DRY and COC.
How about TDD, MVC, ActiveRecord, and block closures?
Eleanor McHugh
> Eleanor McHugh wrote:
>> Because they like Groovy :)
>> It's a good language if the JVM is your only target, and it's still
>> close enough to Java that there's not too much pain working with
>> both.
>
> slow.
Indeed. Although its not that slow compared to other dynamic JVM
languages of similar vintage. JRuby runs rings around it though ;)
Michael T. Richter
On Sat, 2008-07-05 at 04:08 +0900, James Britt wrote:
> If Ruby (with Rails or whatever) is the objectively better choice then
> you should be able to demonstrate that. There has to be something
> tangible, something other than "I like it."
A major supplier of, among other things, medical software recently lost
their security primo who was also a Ruby/Rails advocate. It became
corporate policy to use Java, you see, and absolutely everything they
did had to be done in Java. Never mind that said primo was routinely
able to manoeuvre along with the huge, semi-random requirements changes
without breaking a sweat. Never mind that in his opinion -- keep in
mind that he was hired because he had expertise they didn't have
in-house! -- moving everything over to Java would be a disaster. Never
mind that when the Java group that was battling him tried to replicate
just one snapshot of his code's capabilities in three months of effort
(it took him three weeks to make that snapshot, and that was with
exploring the domain as well as designing and coding -- they had the
design and the domain exploration already done) they "replicated" it by
ignoring all the requirements that he was working under and provided a
service that would require their users to license very expensive
third-party software components to access. Never mind all that:
corporate policy was "everything in Java" so everything had to be in
Java.
I really don't like being cynical, I'm just driven to it by observation
of life. No amount of objective "what's best" will ever reverse a
management decision once it's made. Rationality has no place in
business management, it seems, only ego.
--
Michael T. Richter <ttmri...@gmail.com> (GoogleTalk:
ttmri...@gmail.com)
A well-designed and humane interface does not need to be split into
beginner and expert subsystems. (Jef Raskin)
Eleanor McHugh
My partner in crime Romek has just had a similar experience with a
leading UK bank. They hired him to sort out their PKI, a task which
experience tells us he could have completed in three weeks with Ruby.
Unfortunately they were a Visual C++ house so he spent 6+ months doing
it that way. Now his contract is up and they're switching to Java, so
the whole thing will be redone with bouncy castle. Needless to say,
the PKI is just as unlikely to see the light of day this year if ever.
> I really don't like being cynical, I'm just driven to it by
> observation
> of life. No amount of objective "what's best" will ever reverse a
> management decision once it's made. Rationality has no place in
> business management, it seems, only ego.
Based on my personal experience it's equal parts ego, fear, ignorance
and incompetence. It's a wonder really that we're not still living in
tents and digging for roots for a living.
Martin DeMello
<zn...@cesmail.net> wrote:
> Eleanor McHugh wrote:
>
>> Based on my personal experience it's equal parts ego, fear, ignorance and
>> incompetence. It's a wonder really that we're not still living in tents and
>> digging for roots for a living.
>
> there's wireless Internet.
Living in tents and sniffing for root, perhaps :)
m.
Eleanor McHugh
All the cool kids have yurts for war driving ;)