Help Reqd: Making the Case for Scala in Our Enterprise

[Randall R Schulz]

Hi,

My company's technical advisory board has required us (the engineering
staff and management) to make a detailed evaluation of the suitability
of Scala (vs. Java, having settled on the JVM as the platform) for our
back-end services before we decide which language to use.

There are two fronts on which the case must be made, technical and
business. I think we (the existing advocates of Scala within the
engineering staff) can handle the technical side, but I would like help
making the business case.

In particular, there is a strong feeling that Scala is high-risk. There
is some concern that the language may be a "toy" (apparently in this
context that means one that people like to play with but which is not
suitable for commercial purposes; only one person claims this, and he
likes C...). A more common concern is that we will hamper our already
painfully slow hiring process. In general, the learning curve combined
with the perceived added difficulty to hire amount to obstacles to our
progress in the face of the very demanding few years of development we
now face.

The technical leadership and the engineers themselves are split on the
matter. The leadership has expressed the feeling that they would be a
lot more comfortable going with Scala and justifying that decision if
they had some case studies of Scala use in large-scale businesses.
Twitter's use is a big positive to them but, to contrast, WelthFront
(as a company that is presumed never to need the scaling up we already
know we must do) is not. They want more examples of big systems using
Scala.

Can people point me to businesses that are using Scala in
mission-critical, large-scale (10e5 users and up, e.g.),
architecturally complex architectures?

Other arguments that an engineering or technology VP or a CTO would
understand would also be welcome.

Randall Schulz

[Alec Zorab]

Without knowing anything about your hiring process specifically, but
knowing a lot about the people I interview at my workplace, I'm going
to speculate that you're finding it hard to hire people because there
are very few good candidates. This is going to be invariant with the
language you use. A good applicant will be able to learn scala very
easily if they don't already speak it, and a bad applicant would be
bad anyway!

[Kevin Wright]

On 8 June 2011 15:01, Randall R Schulz <rsc...@sonic.net> wrote:

Hi,

My company's technical advisory board has required us (the engineering
staff and management) to make a detailed evaluation of the suitability
of Scala (vs. Java, having settled on the JVM as the platform) for our
back-end services before we decide which language to use.

There are two fronts on which the case must be made, technical and
business. I think we (the existing advocates of Scala within the
engineering staff) can handle the technical side, but I would like help
making the business case.

In particular, there is a strong feeling that Scala is high-risk. There
is some concern that the language may be a "toy" (apparently in this
context that means one that people like to play with but which is not
suitable for commercial purposes; only one person claims this, and he
likes C...).

So they actively want to chose a language that *isn't* fun and interesting? It probably won't do you any favours in terms of aquisitionretention if you go around telling developers that you've chosen the "boring" option for them...

Being risk-adverse is common amongst the kind of people who make such decisions.

This can also work to your advantage when advocating for Scala, which mitigates a number of risks:

- The risk of obsolescence when your s/w design can't scale to the 1000 core systems we'll have within a decade

- The risk of opportunity-cost, Scala is faster to program in than Java

- The risk of competition moving to something more productive than Java

- The risk of not retaining/attracting quality programmers.

 

A more common concern is that we will hamper our already
painfully slow hiring process. In general, the learning curve combined
with the perceived added difficulty to hire amount to obstacles to our
progress in the face of the very demanding few years of development we
now face.

Everyone I've spoken to about this has said the opposite.  The chance to start working with Scala can act as a very powerful incentive to attract people who otherwise wouldn't consider leaving their current role.  Try posting a role on the mailing lists and see what kind of feedback you get!

Scala is also fun (which makes it look like a toy to some people), this makes your existing staff far less likely to leave.

Your hiring shouldn't suffer.  If anything, the situation will improve following a switch to Scala.

 

The technical leadership and the engineers themselves are split on the
matter. The leadership has expressed the feeling that they would be a
lot more comfortable going with Scala and justifying that decision if
they had some case studies of Scala use in large-scale businesses.
Twitter's use is a big positive to them but, to contrast, WelthFront
(as a company that is presumed never to need the scaling up we already
know we must do) is not. They want more examples of big systems using
Scala.

As already posted on IRC: http://www.scala-lang.org/node/1658?

Typesafe may also be able to provide other references if asked.

I'm personally aware of others, but not at liberty to disclose names.

 

Can people point me to businesses that are using Scala in
mission-critical, large-scale (10e5 users and up, e.g.),
architecturally complex architectures?

Other arguments that an engineering or technology VP or a CTO would
understand would also be welcome.

You could always name-drop the Typesafe advisory board :)

 

Randall Schulz

--
Kevin Wright

gtalk / msn : kev.lee...@gmail.com

mail: kevin....@scalatechnology.com

vibe / skype: kev.lee.wright

quora: http://www.quora.com/Kevin-Wright
twitter: @thecoda

"My point today is that, if we wish to count lines of code, we should not regard them as "lines produced" but as "lines spent": the current conventional wisdom is so foolish as to book that count on the wrong side of the ledger" ~ Dijkstra

[Dennis Haupt]

-------- Original-Nachricht --------
> Datum: Wed, 8 Jun 2011 07:01:53 -0700
> Von: Randall R Schulz <rsc...@sonic.net>
> An: scala-...@googlegroups.com
> Betreff: [scala-debate] Help Reqd: Making the Case for Scala in Our Enterprise

i'll never understand that kind of concern. the difference between a failed project and a running project is the sum of a lot of things. the choice of language is NOT one of them. as long as you don't pick brainfuck, a completely outdated language (pre-OO) or a domain specific one for the wrong purpose (javascript for 3d rendering), you won't fail because of the language.

has anyone ever heard of a project that failed because the wrong language was chosen? i didn't. achieving such a failure is really, really hard.

>
> Can people point me to businesses that are using Scala in
> mission-critical, large-scale (10e5 users and up, e.g.),
> architecturally complex architectures?

take a random java project and claim it can be converted almost 1:1 into scala. since your claim is true, why isn't that a proof?

[Chris Marshall]

Randall - the Guardian (www.guardian.co.uk) use scala as a backend for some of their website content. I'm guessing that they would claim to having 10E5 users, if maybe not all at the same time!

I'm with the others on the hiring front; I think it's just a non-argument. A great Java developer will become a great Scala developer with a high degree of certainty. Certainly they will be *as* proficient in Scala as in Java and I would argue that you can replace "as" with "more" in this sentence. The main question is whether you will find your field of candidates whittled down because some great Java devs will have no wish to learn Scala. I think that this is a possibility, sure; but I suspect that these will be more than compensated for by developers who would specifically *not* consider a pure Java shop.

Chris

[Craig Tataryn]

+1

[Justin du coeur]

On Wed, Jun 8, 2011 at 11:20 AM, Chris Marshall <oxbow...@hotmail.com> wrote:

The main question is whether you will find your field of candidates whittled down because some great Java devs will have no wish to learn Scala. I think that this is a possibility, sure; but I suspect that these will be more than compensated for by developers who would specifically *not* consider a pure Java shop.

This is a key point.  To some degree, it depends on what you're trying to hire.  The simple truth is that finding workaday scutwork programmers *is* easier for Java, and will be so for a few years yet.  (I expect that to change, but it needs some educational evolution.)

But high-flyers are getting suspicious of pure Java shops -- Java is looking so old and obsolete at this point that top-notch programmers are starting to avoid it.  So Scala is a good pull for drawing in the best and the brightest, which is usually what you care about most.  

It's relatively easy to find low-level programmers, and they will need training anyway.  The top-notch folks are the ones who are really hard to hire, and a cutting-edge platform is a good way to get their attention...

[Kevin Wright]

+1

[Randall Schulz]

Hi, Folks,

Thanks for the responses so far. (Keep 'em coming!)

On hiring: Yes, we have pretty high standards and we interview many to hire one. (And phone screen many to interview one.)

Those of us already using Scala and advocating for a much wider role for it have made the argument (somewhat successfully, I think) that people who meet our general software engineering standards would not be put out by having to learn Scala. But we are also looking for specific examples of people's experiences in such a circumstance.

Thanks for the Guardian reference, though I see now it's on http://www.scala-lang.org/node/1658, too.


Randall Schulz

[Peter White]

Hi, All,

I'm with Randall - any developer who gets through our interview process at Box should be more than capable of becoming productive with Scala rather quickly. I believe the resistance faced by the pro-Scala camp is the perception that:

1) The pool of great programmers interested in programming in Java is > 100x larger than the pool of great programmers interested in programming in Scala. This *isn't* a quantity vs. quality issue - we need to hire on a massive scale, and the people we hire need to become productive in whichever language we choose ASAP. Frankly speaking, I think the high-bar of our interview process is more of an impediment to meeting our hiring goals than the preconceived notion that we'll be able to find enough great Java developers but we'll fail to find enough great developers who can learn enough Scala to be productive quickly. It's also true that *some* great Java developers won't be willing to switch to Scala.

2) The Scala tool-chain (edit, build, compile, debug, deploy, etc) is less mature than Java's - aside from Java IDEs currently having better refactoring support, I consider both languages to be equivalent do to the fact that it's all JVM-bytecode in the end.

3) There's some concern that Scala could be just another fad that loses steam and gets abandoned after some undetermined period of time. The line of thought as been it's much easier to migrate to Scala after it becomes more of a proven quantity, but that we're going to be in a very bad position if we need to revert from Scala to Java for any reason.

Just to reiterate what Randall said, case studies of companies that have successfully made the switch to Scala for their back-end development as well as case studies of companies that have rapidly grown (10 -> 30+, for example) Scala development teams would be extremely helpful with respect to making our case that not only is Scala a "safe bet", but it's also a *Competitive Advantage*.

Shameless Plug - I apologize in advance if job postings are considered inappropriate in this group... I'm looking to hire a Software Engineer in the "Tools and Frameworks" group @ Box, and we HAVE standardized on the use of Scala within this team. Please see the "Software Engineer - Tools and Frameworks" job description under our current job openings page for more info and to apply if you're interested.

Thanks!

Peter

[Kevin Wright]

If posting Job information, can I please ask that you mention where you're based... Either that, or state that you're willing to take on telecommuters.

Anything else is just teasing!

[Steven Obua]

Just click on the link :-)

[Peter White]

Sorry, Kevin,

Box is located in Palo Alto, CA. Due to the highly collaborative nature of our organization (communication with other team members as well as our end-users), telecommuting is not an option. We have pretty flexible work hours, and telecommuting is possible when an employee is stuck at home for various reasons on occasion, but we're unable to accomodate candidates who need to consistently telecommute N days a week.

- Peter

[Kevin Wright]

Oh, I'm in the UK - so even telecommuting would be a bit extreme for me, you have time zones and a language barrier to contend with as well.

My appeal was more of a general one for anyone else who posts job info :)

[Sergei]

These three perceptions that Peter listed were precisely the top three
I had to overcome at my current place of employment. Eventually I
achieved partial victory (those who want to use Scala can use it), yet
Java still dominates the code base and is the default for new
projects. The technique I used for convincing the upper management was
a traditional mix of technical superiority + business benefits
illustrations. Examples of Twitter, Foursquare, and especially
LinkedIn using Scala in production were significant influencers as
well.

Couple more insights about the Scala's current competition that I
encountered in the process:

(1) In regards to Java programmers, Scala competes not so much with
Java the language as with programming patterns and idioms that
experienced developers built around Java. In other words, good Java
developers do not program in pure Java, they program in multiple
domain-specific languages built with Java, usually expressed as
prescribed patterns of calls and call-backs using domain-specific Java
libraries. Some of these DSLs and libraries are specific to a
particular developer, team, or a company and influence significant
part of their perceived self-worth, job security, competitiveness, and
other psychologically desired values.

(2) I think that an experienced, accomplished, deeply knowledgeable
Java expert is perhaps the worst target of conversion to Scala. It is
based not only on my recent experiences, but also on experiences from
the era when Java was to C++ what Scala is to Java today (not precise
analogy of course). In 1996 I worked as a senior architect on a system
that was regarded by Sun as the largest business Java project in
existence at that time. It was not a particularly smooth going - e.g.
JSP and JDBC were not yet invented when the project was started, and
development tools were rough around the edges. Yet the project was
ultimately successful and the system is still around, generating tons
of revenue. The developers we hired were either (a) language polyglots
and early adopters like me or (b) kids straight out of college, with
minds not yet contaminated with C++ "advanced patterns".

Back then as today, the deepest experts in the currently dominant
programming language paradigm were either the last to convert or would
never convert at all. So, to address the perceived lack of Scala
programmers, you may use non-traditional approaches - for instance,
shamelessly and aggressively pursuing programming language polyglots
currently employed not only in Java, but also in Python, Ruby, .Net
etc. shops. Search not only for established experts and top college
graduates, but also for smart, energetic kids who, due to
circumstances beyond their control, could only make it to a community
college - e.g. one of the best Java programmers we had back in 1996
only had a liberal arts degree from a tiny college in the middle of
nowhere.

Hope this helps,
Sergei.

> Frameworks" job description under our current job openings<http://www.box.net/company/careers/job_openings>page for more info and to apply if you're interested.
>
> Thanks!
> Peter

[Randall R Schulz]

On Wednesday June 8 2011, Randall R Schulz wrote:
> ...

Thanks everyone for the replies, including the private ones.

This has been very helpful.

Randall Schulz

[martin odersky]

On Wed, Jun 8, 2011 at 7:28 PM, Peter White <pwh...@gmail.com> wrote:

Hi, All,

I'm with Randall - any developer who gets through our interview process at Box should be more than capable of becoming productive with Scala rather quickly. I believe the resistance faced by the pro-Scala camp is the perception that:

1) The pool of great programmers interested in programming in Java is > 100x larger than the pool of great programmers interested in programming in Scala. This *isn't* a quantity vs. quality issue - we need to hire on a massive scale, and the people we hire need to become productive in whichever language we choose ASAP.

To give some quantitative information on this, here's the latest job trends on indeed.com:

http://www.indeed.com/jobtrends?q=scala%2C+clojure%2C+erlang%2C+f%23&l=

The statistics show two things:

1) Scala jobs are growing very fast, and at a steady rate. There's a clear hockey stick curve. This should address any concerns that this is a toy or a fad.

2) In absolute terms, Scala job ads amount to 0.6% of Java jobs ads in the US. In some other markets such as the UK that percentage is closer to 1%.
At current growth rates this would mean ~2.5% of the Java market in 2 years.
That's a low bar - we formed Typesafe with the aim of accelerating that growth rate.

If you make the assumption that job supply is roughly in sync with job demand, and make the additional assumption that Scala appeals particularly to high-end programmers, then even today you should do considerably better than 1 great Scala dev per 100 great Java devs.

Cheers

 -- Martin

[Chris Marshall]

Martin - 

Relevant to this discussion would be percentage of job postings that remain unfilled after a specified time period (e.g. 3 months). If scala jobs are disproportionately hard to fill, then that is ammunition against Scala. If not, then that is an effective riposte to a common argument. Do you have or know of any figures for this?

After all, the growth rate might just be a side effect of old postings staying around because the job is still available (although that's unlikely).

Chris

---

From: martin....@epfl.ch
Date: Thu, 9 Jun 2011 11:51:11 +0200
Subject: Re: [scala-debate] Re: Help Reqd: Making the Case for Scala in Our Enterprise
To: scala-...@googlegroups.com

[Robert Wills]

On Thu, Jun 9, 2011 at 10:51 AM, martin odersky <martin....@epfl.ch> wrote:

If you make the assumption that job supply is roughly in sync with job demand, and make the additional assumption that Scala appeals particularly to high-end programmers, then even today you should do considerably better than 1 great Scala dev per 100 great Java devs.

Assuming that developers with Scala experience are more likely to be attractive than most Java developers it strikes me that companies adopting Scala should consider bypassing our friends the recruitment consultants.  Some company in London is adopting Scala for a project and I had 3 calls from separate recruitment consultants yesterday -- I don't understand why that company doesn't just post to the scala-user list or go to Typesafe or one of the other scala consulting companies?   Is it worth having a Scala Jobs section on scala-lang.org or somewhere similar?

-Rob

[Kevin Wright]

You're not alone, I think many of us have been getting the recruiters and the LinkedIn invites.  The way the roles are being described, I get the impression that there are two distinct companies out there who have just started looking for Scala devs in a serious way.

[martin odersky]

On Thu, Jun 9, 2011 at 12:14 PM, Chris Marshall <oxbow...@hotmail.com> wrote:

Martin - 

Relevant to this discussion would be percentage of job postings that remain unfilled after a specified time period (e.g. 3 months). If scala jobs are disproportionately hard to fill, then that is ammunition against Scala. If not, then that is an effective riposte to a common argument. Do you have or know of any figures for this?

I don't have any data on that aspect. I agree it would be highly relevant. If somebody knows how to extract this I would be very interested in the result.

Thanks

 -- Martin
 

After all, the growth rate might just be a side effect of old postings staying around because the job is still available (although that's unlikely).

Chris

---

From: martin....@epfl.ch
Date: Thu, 9 Jun 2011 11:51:11 +0200
Subject: Re: [scala-debate] Re: Help Reqd: Making the Case for Scala in Our Enterprise

To: scala-...@googlegroups.com

On Wed, Jun 8, 2011 at 7:28 PM, Peter White <pwh...@gmail.com> wrote:

Hi, All,

I'm with Randall - any developer who gets through our interview process at Box should be more than capable of becoming productive with Scala rather quickly. I believe the resistance faced by the pro-Scala camp is the perception that:

1) The pool of great programmers interested in programming in Java is > 100x larger than the pool of great programmers interested in programming in Scala. This *isn't* a quantity vs. quality issue - we need to hire on a massive scale, and the people we hire need to become productive in whichever language we choose ASAP.

To give some quantitative information on this, here's the latest job trends on indeed.com:

http://www.indeed.com/jobtrends?q=scala%2C+clojure%2C+erlang%2C+f%23&l=

The statistics show two things:

1) Scala jobs are growing very fast, and at a steady rate. There's a clear hockey stick curve. This should address any concerns that this is a toy or a fad.

2) In absolute terms, Scala job ads amount to 0.6% of Java jobs ads in the US. In some other markets such as the UK that percentage is closer to 1%.
At current growth rates this would mean ~2.5% of the Java market in 2 years.
That's a low bar - we formed Typesafe with the aim of accelerating that growth rate.

If you make the assumption that job supply is roughly in sync with job demand, and make the additional assumption that Scala appeals particularly to high-end programmers, then even today you should do considerably better than 1 great Scala dev per 100 great Java devs.

Cheers

 -- Martin

--
----------------------------------------------
Martin Odersky
Prof., EPFL and CEO, Typesafe
PSED, 1015 Lausanne, Switzerland
Tel. EPFL: +41 21 693 6863
Tel. Typesafe: +41 21 691 4967

[Steven Obua]

Given the recent job discussion here, I thought this might be a good opportunity to post my cv, as I am looking for a Scala or ML job :-)
Feel free to email me if you are considering to hire me as part of your (not too large) team. Freelance would be interesting, too.

Any advice on how to best go looking for a Scala job is appreciated, too!

- Steven

PS:
My nationality is German, so US jobs seem to be problematic because of VISA stuff and all that. Nevertheless, I would like to work in the US, preferably Bay Area or NY. But I am pretty flexible in that respect, the most important criterium is the actual job.

[Dennis Haupt]

you should add some more proofs of your skill.

for example, i did:

+ 3d engine (youtube, search for "hhexen jmonkey")
+ an online shop (like amazon, just a *little* smaller ;))
+ ftp applet
+ a starcraft bot
+ some more stuff

-------- Original-Nachricht --------
> Datum: Thu, 9 Jun 2011 13:37:25 +0200
> Von: Steven Obua <steve...@googlemail.com>
> An: scala-debate <scala-...@googlegroups.com>
> Betreff: [scala-debate] Looking for Scala Job

[Goddard Jiri]

@Dennis PLease look into the CV as I did and then look up the iPhone apps in Appstore for evaluation :)

Regards, Jiri

--
web: http://www.dredwerkz.cz

blog: http://dr3dwerkz.blogspot.com/

code: https://github.com/g0dd4rd
group: http://groups.google.com/group/dr3dwerkz
music: http://profile.ultimate-guitar.com/g0dd4rd/
twitter: http://twitter.com/#!/g0dd4rd
profile: http://www.google.com/profiles/g0dd4rd
icq: 218 659 431

[Dennis Haupt]

i hereby change "add proof" to "layout the written proof so that the quickly reading human resource guy can't overlook it"

-------- Original-Nachricht --------
> Datum: Thu, 9 Jun 2011 14:35:52 +0200
> Von: Goddard Jiri <g0d...@googlemail.com>
> An: Dennis Haupt <h-s...@gmx.de>
> CC: Steven Obua <steve...@googlemail.com>, scala-...@googlegroups.com
> Betreff: Re: [scala-debate] Looking for Scala Job

> <http://www.dredwerkz.cz>code: https://github.com/g0dd4rd

[Philippe Lhoste]

On 08/06/2011 20:23, Kevin Wright wrote:
> Oh, I'm in the UK - so even telecommuting would be a bit extreme for me, you have time
> zones and a language barrier to contend with as well.
>
> My appeal was more of a general one for anyone else who posts job info :)

I fully agree, this information is often missing, or not made prominent, which is a bit
surprising given that I look at these job offers from Internet, accessible worldwide.
Of course, I suppose that when not said otherwise, the job offer is not for telecommuting,
ie. it is for local job, but even this information is sometime missing from the job offer,
or you have to go elsewhere on the Web site to find it.
In the case of the job offer on this thread, I had to look attentively to find a location
(US, so not for me...).

Just offering my experience/opinion of foreigner (from non-French PoV) from somebody not
searching very actively, but wanting to find a cool job, preferably in the Scala world
(like lot of us (readers of the ML) not already there, I suppose).

Opportunities in France aren't so numerous, yet, so telecommuting is seductive for
somebody not wanting to expatriate with all the family... :-)

--
Philippe Lhoste
-- (near) Paris -- France
-- http://Phi.Lho.free.fr
-- -- -- -- -- -- -- -- -- -- -- -- -- --

[Jim Powers]

Couple of points:

Firstly, Martin, in your keynote at Scala Days 2011 you had the outside.in logo on the slide about what companies are using Scala.  It was true that we used Scala, but outside.in was acquired at the end of February by AOL and merged into the Patch group.  What we did at outside.in was hyper-local news aggregation and our data collection and internal and public APIs are written in Scala.  The aggregation system of outside.in is alive and well at Patch so you can substitute the Patch logo for outside.in.  To add our voice to the choir: Patch is also looking for Scala developers ;-).

Secondly, a few comments on our hiring experiences.  Early on (~2.5-3 years ago when we started using Scala) it was get smart+motivated people with a good programming track-record and train, few alternatives presented themselves.  Today it's still mostly that, but now the local community (in NYC and surrounding areas) of Scala developers is growing and more hiring is happening between members of that community, but the community is no where large enough to meet demand.  Everyone in this community pretty-much knows everyone else and we all know we're poaching from one-another.  Another recent development that supports Martin's info on increased job postings is that more and more of the people we are interviewing have heard of Scala and an increasing number have either toyed with it or built something small in their personal time.  The message here is that more people have at least some learning under their belt which helps when discussing Scala at the interview.  To some of the points that have been brought up regarding the "more elite" programmers being drawn to Scala, I cannot back up this claim scientifically, but it feels true, however, this is a double-edged sword.  The problem arises when you try to hire for budgeted job slots.  We have had to turn down some clearly "elite" developers because there was no chance we could afford them. As a fraction of the population the "expensive elites" are going to be disproportionally represented compared to such a population in the Java talent pool.  We are still years away IMHO to seeing a nice population of entry-level Scala developers on the market, to address some of that it would be nice to see some schools replace Java with Scala in their curriculum (at least starting the process of fixing what Java hath broken ;-) ).  At the entry and mid-levels I would think that you just have to work with the brightest minds you can find and train.

--

Jim Powers

[martin odersky]

Thanks for the correction, Jim. I'll change the logo to Patch. I see some Scala courses springing up at universities, but overall progress is still a bit slow.

 -- Martin

[Justin du coeur]

On Thu, Jun 9, 2011 at 9:47 AM, Jim Powers <j...@casapowers.com> wrote:

We are still years away IMHO to seeing a nice population of entry-level Scala developers on the market, to address some of that it would be nice to see some schools replace Java with Scala in their curriculum (at least starting the process of fixing what Java hath broken ;-) ).  At the entry and mid-levels I would think that you just have to work with the brightest minds you can find and train.

Okay, time to return to my periodic hobby-horse: is there yet a decent entry-level book / online course for Scala?  Martin's pedagogical outline a few months ago was a great start at breaking things down into nicely bite-size chunks; has anybody reified that into a course yet?

Seriously: this is a project Typesafe might want to think about pursuing seriously.  I suspect that a really solid "Entry-Level Scala" course online would do wonders for adoption, and have some nice side-effects: it would provide folks with a gut-level sense that you don't have to be a CS whiz in order to "do" Scala; it would show the business community that their needs are being addressed; etc.

[Jim Powers]

Learn you a Scala for great good? ;-)

--

Jim Powers

[Justin du coeur]

Being in the middle of LYaHFGG myself -- yeah, something like that.

But more importantly: Martin did a nice job of breaking down the language features in terms of what *kind* of programmer is likely to care about them.  That nicely illustrated that a lot of the "hard" features aren't necessarily crucial to the life of the average low-level apps programmer.  A course that was aimed specifically at *that persona* (which, let's get real, probably accounts for about 80% of all programmers) would be really useful.  It would probably teach just a subset of the language (and yes, something in the breezy style of Learn You a Haskell would be brilliant), with tantalizing "for more information" hints about the cooler stuff.

I'm thinking specifically the "foot in the door" course, that gives you enough to be productive *specifically* as a basic apps programmer, even if you don't know all the ins and outs.  That would likely be relatively short and easy, and would go a good ways to make Scala look less intimidating.  (And meanwhile, the "for more information" hints are the candy to encourage people to start broadening their horizons...)

[Peter C. Chapin]

 

I plan to use Scala in my Programming Languages course this coming fall as a vehicle for talking about various programming language features. The students are mostly experienced with Java (some C and C++ also). While focusing on a single language in a course called “Programming Languages” may seem wrong, I like giving the students enough exposure with a particular language to get them comfortable with it. My experience with the course in the past is that if one just talks about language topics in isolation (not in the context of a single language) the students just don’t get the point.

 

I’m using Scala because

 

1)      I like it.

2)      It offers a rich collection of features in a nicely integrated package.

3)      It’s extremely practical in the real world and I believe my students will appreciate that.

 

I think the last point is very important. My student audience will want to know, “can this help me get a job?” With Scala I feel like I can confidently say, “yes.” In fact I’d like to think having some Scala experience on their resume will set them apart from the Java minions that are out there. Alas the other faculty here mostly roll their eyes and tolerate me when I start talking about Scala. Ah well.

 

In any case there is definitely some Scala creeping into the college classroom. I’m doing it.

 

Peter

 

 

From: scala-...@googlegroups.com [mailto:scala-...@googlegroups.com] On Behalf Of Justin du coeur
Sent: Thursday, June 09, 2011 12:47
To: Jim Powers
Cc: scala-...@googlegroups.com
Subject: Re: [scala-debate] Re: Help Reqd: Making the Case for Scala in Our Enterprise

 

On Thu, Jun 9, 2011 at 9:47 AM, Jim Powers <j...@casapowers.com> wrote:

[Marc Edwards]

Seems the thread has diverged a little awry from the original post.

So, in scanning this thread, the production Scala projects mentioned include: (1) Twitter, (2) a group at Box.net, (3) Patch Network.

Does this look correct?

[Francois]

Well, and all project on the page: http://www.scala-lang.org/node/1658
So it is not really correct.

-- 
Francois ARMAND
http://fanf42.blogspot.com
http://www.normation.com

[Randall R Schulz]

The "group at Box.net" (of which I am a member) is where we're trying to
decide whether to take on Scala or Java as our JVM language as our
back-end services language. When the topic addressed acquisition of
talent, my supervisor at Box, Peter White, joined in to describe our
hiring practices.

Randall Schulz

[James Edwards]

So, if we wanted to prioritize the risks to adopting Scala, who would
agree with this prioritized list?

1) Lack of skilled programming base.
2) Immaturity of tools
3) Immaturity of compiler?
4) Robustness of open source library support? (Since all Java
libraries can be used, it seems that the available library support is
a non-issue)

The bottom line from an executive's perspective is whether or not the
programming team can deliver the project on schedule, budget. In
addition, can they maintain the environment and enhance it with the
underlying programming base. There can be NO CONDITIONS under which
the engineering team comes back and says, "hey, we can't get our work
done, because the compiler or tools are screwing us up." Then the
question is who is going to fix this and how much will it cost? And,
then comes the search for a new engineering team!!

-M

On Jun 8, 10:01 am, Randall R Schulz <rsch...@sonic.net> wrote:
> Hi,
>
> My company's technical advisory board has required us (the engineering
> staff and management) to make a detailed evaluation of the suitability
> of Scala (vs. Java, having settled on the JVM as the platform) for our
> back-end services before we decide which language to use.
>
> There are two fronts on which the case must be made, technical and
> business. I think we (the existing advocates of Scala within the
> engineering staff) can handle the technical side, but I would like help
> making the business case.
>
> In particular, there is a strong feeling that Scala is high-risk. There
> is some concern that the language may be a "toy" (apparently in this
> context that means one that people like to play with but which is not
> suitable for commercial purposes; only one person claims this, and he
> likes C...). A more common concern is that we will hamper our already
> painfully slow hiring process. In general, the learning curve combined
> with the perceived added difficulty to hire amount to obstacles to our
> progress in the face of the very demanding few years of development we
> now face.
>
> The technical leadership and the engineers themselves are split on the
> matter. The leadership has expressed the feeling that they would be a
> lot more comfortable going with Scala and justifying that decision if
> they had some case studies of Scala use in large-scale businesses.
> Twitter's use is a big positive to them but, to contrast, WelthFront
> (as a company that is presumed never to need the scaling up we already
> know we must do) is not. They want more examples of big systems using
> Scala.
>
> Can people point me to businesses that are using Scala in
> mission-critical, large-scale (10e5 users and up, e.g.),
> architecturally complex architectures?
>
> Other arguments that an engineering or technology VP or a CTO would
> understand would also be welcome.
>
> Randall Schulz

[Jim Powers]

On Fri, Jun 10, 2011 at 10:53 AM, James Edwards <jmarce...@gmail.com> wrote:

The bottom line from an executive's perspective is whether or not the
programming team can deliver the project on schedule, budget.  In
addition, can they maintain the environment and enhance it with the
underlying programming base.  There can be NO CONDITIONS under which
the engineering team comes back and says, "hey, we can't get our work
done, because the compiler or tools are screwing us up."  Then the
question is who is going to fix this and how much will it cost?  And,
then comes the search for a new engineering team!!

There are cases where Scala "won't work"[1] such as generating Java-style enums.  Typically this only really matters if there is some Java code-base where you need to supply Java enum values.  In the couple of cases where I've encountered this banging out a Java class to get me the enums I need is all it took to get over this hurdle.  I'm sure that there are other cases (wacked-out overlaoded class constructors that expose disjoint functionality depending on what constructors are used [I've not encountered such a case, just theorizing], might be another case).  But the point really is: if Scala is literally stopping you from moving forward and the greatest minds in the Scala universe cannot get you through this impediment, there still is always plain-old Java to get you back on your way.

 

--

Jim Powers

[1] In this particular example not supporting Java-style enums (which I completely agree with) is a language design decision.

[Erik Engbrecht]

There are cases where Scala "won't work"[1] such as generating Java-style enums.  Typically this only really matters if there is some Java code-base where you need to supply Java enum values.  In the couple of cases where I've encountered this banging out a Java class to get me the enums I need is all it took to get over this hurdle.  I'm sure that there are other cases (wacked-out overlaoded class constructors that expose disjoint functionality depending on what constructors are used [I've not encountered such a case, just theorizing], might be another case).  But the point really is: if Scala is literally stopping you from moving forward and the greatest minds in the Scala universe cannot get you through this impediment, there still is always plain-old Java to get you back on your way.

You need to consider more complex scenarios.  For example, let's say considerable time is put into designing and coding a module that makes heavy use of the Scala XML libraries, and late in the game a memory leak (or perhaps something that isn't technically a "leak," but highly suboptimal and leads to memory being held much longer than it should be) is discovered that causes the system to crash and burn in certain high-load circumstances (I'm not saying there is such a problem, this is purely hypothetical).

Going back to plain-old Java would be expensive because it would require a re-write.  Chances are the library can be fixed, but that is likely to present a choice between running an unreleased version of Scala either in the form of patching the version currently being used or upgrading to a non-release build.  These aren't pleasant choices.  Also, if you end up developing the patch yourself there's no guarantee it will ever make it into the main distribution.

Of course this risk exists with any technology, but it's particularly pronounced when one is an early adopter, or when one is pushing a technology farther than others have done.

-Erik

[Nikita Ivanov]

I'm talking to our potential customer about Scala almost weekly. Here
are the observations (everything is IMHO):

1. Professional Scala devs don't exist (I mean quite literally). Good
luck even finding the resume. You need to hire good Java dev and give
him week or two to learn Scala and go from there. If he/she smart
enough - Scala can be picked up in ~1-2 months (including, of course,
"proper" Scala development + all esoteric stuff like type programming,
etc.)

2. Many large enterprise customers will reject Scala today b/c "my
guys in Bangalore/Beijing" just finally started to produce decent Java
code. No way they are going to throw Scala on them in any foreseeable
future.

3. On-site, small teams are the best for adoption of Scala. This has
been proven time after time.

4. Cultural differences are abound... Europeans are much more in tune
with FP (due to education) and therefor Scala doesn't seem so foreign
as in USA/Indian/China/Russia. Companies that have significant
European workforce can be more friendly to the idea.

Hope it helps,
--
Nikita Ivanov
GridGain Systems

[Robert Wills]

On Mon, Jun 13, 2011 at 7:43 PM, Nikita Ivanov <niv...@gridgain.com> wrote:

if he/she smart

enough - Scala can be picked up in ~1-2 months (including, of course,
"proper" Scala development + all esoteric stuff like type programming,
etc.

2. Many large enterprise customers will reject Scala today b/c "my
guys in Bangalore/Beijing" just finally started to produce decent Java
code. No way they are going to throw Scala on them in any foreseeable
future.

There's an assumption here (repeated many times in other posts) that I 

think needs challenging or at least questioning: namely, that you can need

to start with a GOOD Java developer if you want to get a Scala developer.

I think it is very difficult to be a good Java developer. My memory of it is that 

it was extraordinarily easy for everything to turn to mush after a few cycles of 

development -- I wasn't the worst Java developer but I don't think I was a good

one.  With Scala on the other hand I find it relatively to keep a project reasonably

non-ugly.

Maybe part of the sales pitch for Scala ought to be that it may help you to 

uncover hidden talent in your organisation.  All those people worn down by what 

Bruce Eckel called the 'ceremony' of Java programming might actually get interested

in programming again.  Those guys in Bangalore and Beijing might actually turn

out to produce better Scala code.

 

[James Edwards]

http://bostinnovation.com/2011/06/16/scalas-popularity-on-the-rise-in-boston-sbt-for-android/

May have already been posted, but just in case...

[Razvan Cojocaru]

I had some arguments here:
http://www.slideshare.net/razvanc/why-scalaexecutive-overview this was meant
as an executive overview for our CTO...

-----Original Message-----
From: scala-...@googlegroups.com [mailto:scala-...@googlegroups.com]
On Behalf Of Randall R Schulz
Sent: June-08-11 10:02 AM
To: scala-...@googlegroups.com
Subject: [scala-debate] Help Reqd: Making the Case for Scala in Our
Enterprise