Language Popularity & Network Effects; Ruby & Mongrel
[1/15] from: edoconnor::gmail::com at: 20-Nov-2007 9:57
A few interesting reads I stumbled upon recently:
1. You Work in a Fashion Industry
...if you want to introduce a language, you don't concentrate on
making it a good language, you try to persuade the herd of
programmers, PMs and tool vendors that your language is the Next Big
Thing. The important point here is not how much the language will do
for productivity, quality and cost, it is to create the perception
that everyone else thinks that this language will be the next big
thing.
There are two ways to do this....
http://rubygarden.org/ruby/page/show/YouWorkInAFashionIndustry
2. Interview with Zed Shaw, the author of Mongrel, the ruby webserver.
What I found most interesting is how motivated the community is, and
how they worked to fix some of ruby's problems with arrays and
threads.
http://www.infoq.com/interviews/ruby-zed-shaw
Regards,
Ed
[2/15] from: carl::cybercraft::co::nz at: 21-Nov-2007 8:28
From the fashion industry one...
Trained staff in a minority language are going to be rare. This does not necessarily
make them more expensive (nobody else wants them), but it does make recruitment much
harder and more uncertain. Alternatively you have to train all your existing people in
the new language. And for Functional Languages its not just another syntax, its a whole
new way of thinking.
Two marks against REBOL there.
Still, he did suggest a road to popularity - find a niche and own it. What's a niche
REBOL could create/take over?
-- Carl Read.
On Tuesday, 20-Novenber-2007 at 9:57:25 Ed O'Connor wrote,
[3/15] from: rasmussen:bryan::gmail at: 21-Nov-2007 8:38
>This does not necessarily make them more expensive (nobody else wants them)
This is true in theory but in actuality I think it often does make
them more expensive because most n minority languages that are
established as minority languages that wont disappear tomorrow are
already niche languages, and to get that niche support companies often
have to pay more.
frankly I am at a loss as to what Rebol's niche could be.
Cheers,
Bryan Rasmussen
[4/15] from: petr:krenzelok:seznam:cz at: 21-Nov-2007 9:00
Carl Read napsal(a):
> >From the fashion industry one...
>
> "Trained staff in a minority language are going to be rare. This does not necessarily
make them more expensive (nobody else wants them), but it does make recruitment much
harder and more uncertain. Alternatively you have to train all your existing people in
the new language. And for Functional Languages its not just another syntax, its a whole
new way of thinking."
>
> Two marks against REBOL there.
>
> Still, he did suggest a road to popularity - find a niche and own it. What's a niche
REBOL could create/take over?
>
Flash/Flex area - small, agile, distributed self-served PLATFORM, not
just a language. There is long term plan - Altissimo - mixture of
ViewTop, IOS, Altme, platform for small distributed apps. Another niche?
Embedded area? We will see ...
Petr
[5/15] from: carl::cybercraft::co::nz at: 21-Nov-2007 12:28
On Wednesday, 21-Novenber-2007 at 9:00:33 Petr Krenzelok wrote,
>Carl Read napsal(a):
>> >From the fashion industry one...
<<quoted lines omitted: 13>>
>just a language. There is long term plan - Altissimo - mixture of
>ViewTop, IOS, Altme, platform for small distributed apps.
An online desktop then, but with no need of a central server owned by others?
A Google for "online desktop" turns up plenty of attempts, but I suspect all of them
are based on central-servers. In other words, designed for lock-in. (No time to test
them. I looked at the offerings a year or so back, and some looked promising in that
they were relatively responsive, while others were horribly slow.)
But won't those who don't like the central server approach also be the ones who expect
the alternative to be completely open-source? (REBOL would also be great for a distributed
social network, except those who don't like the current social networks being central-server
based are not going to accept an alternative with software that isn't open-source.)
But yes, technically, REBOL would be great in both those areas. And ought to able to
attract those who would like an online desktop/platform and don't care about how it's
done. (I'm pretty sure a distributed social-network's a non-starter though.)
-- Carl Read.
[6/15] from: jblake::arsenaldigital::com at: 21-Nov-2007 8:37
SOA implementation.
I know it's an "architecture" but to be able to enable a service on
almost any server you have; and then write and deploy code to use that
service from any server you have; and be able to use the code from any
server you have; all with loading an app that's say 1 mg max?
I cant wait for R3 so I can use the "services", dbase access and cli
part.
Then I'll stop using Perl.
John
[7/15] from: ale870:gma:il at: 21-Nov-2007 17:23
Hello,
central servers are a good way, for BIG companies, to "control" what you do,
to control you and your PC (make publicity and money!).
The same target are Ajax, GWT, Google OS, Flex, Adobe Air, etc...
On Nov 21, 2007 2:37 PM, John Blake <jblake-arsenaldigital.com> wrote:
> SOA implementation.
> I know it's an "architecture" but to be able to enable a service on
<<quoted lines omitted: 60>>
> To unsubscribe from the list, just send an email to
> lists at rebol.com with unsubscribe as the subject.
--
//Alessandro
http://sguish.wordpress.com
http://laccio.wordpress.com
[8/15] from: carl::cybercraft::co::nz at: 21-Nov-2007 20:33
On Wednesday, 21-Novenber-2007 at 17:23:27 Alessandro Manotti wrote,
>Hello,
>
>central servers are a good way, for BIG companies, to "control" what you do,
>to control you and your PC (make publicity and money!).
>The same target are Ajax, GWT, Google OS, Flex, Adobe Air, etc...
It's not just 'big' companies though, but any group who has control of those servers
when you can't move your software/data elsewhere without losing some or a lot of its
functionality.
-- Carl Read.
[9/15] from: edoconnor:gma:il at: 21-Nov-2007 15:53
Interesting discussion. I like the notion of PHP finding its niche
with serving web pages.
I'm not sure what REBOL's niche would be. REBOL was designed as a
messaging language. I think I'd be struck down to hear someone say,
Boy, you need to get your hands on a good, solid messaging language!
Or, likewise, "Where can I find a dynamic, highly-reflective
metaprogramming language when I need one." :^)
I think that REBOL's strong point is it's convenience. I like the fact
that I can fire up the interpreter on Windows or OS X and use it to
manage my local data files with minimal fuss. Or that I can parse &
extract content from web pages quickly. In this personal
productivity/end-user programming space, I'd like to see stronger ways
of managing/launching automated scripts, like a cron-job on the *nix
platforms.
The areas where I find REBOL is not so simple is in managing errors,
networking, encryption, xml and building DSLs. I'm not saying the
power isn't there, just that you need to have a good deal of expertise
in these areas to leverage these features. Oddly, many of the
aforementioned features might be considered central to the idea of
messaging
. If REBOL could make these things much more accessible and
simple, and co-opt more of the SOA terminology, I think RT could enjoy
more acceptance as a quick & dirty scripting language for plumbing
enterprise architectures (as well as public ones, amazon S3, etc.).
REBOL needs to enable the type of programming experience and end-user
apps that people are not getting from current scripting languages and
tool-sets. See "We have lost control of the apparatus" for more on
anticipating the needs of future users:
http://weblog.raganwald.com/2007/09/we-have-lost-control-of-apparatus.html
Regards
[10/15] from: ale870::gmail at: 21-Nov-2007 22:43
Ed, I think you centered the problem!
Rebol is great for communication, so maybe, it could increase its power just
in this way:
1) implement more protocols (e.g.: ldap, snmp, protocol for M$ Outlook, XML,
SOAP, JSON, etc...)
2) increment quality of client and GUI
A lot of "standard" scripting languages exist, but when I use Rebol it is a
joy for http handling, html parsing (but maybe, it is still too much
complex. I think we need something like DOM navigation, but using Rebol
power like native datatypes, blocks, etc...).
Rebol has a good area where it can grow: communication, console piping,
applications piping, etc...
This area has almost no competitor, infact every other scripting language
has not so much flexibility that Rebol has!
On Nov 21, 2007 9:53 PM, Ed O'Connor <edoconnor-gmail.com> wrote:
> Interesting discussion. I like the notion of PHP finding its niche
> with serving web pages.
<<quoted lines omitted: 28>>
> To unsubscribe from the list, just send an email to
> lists at rebol.com with unsubscribe as the subject.
--
//Alessandro
http://sguish.wordpress.com
http://laccio.wordpress.com
[11/15] from: tim-johnsons::web::com at: 21-Nov-2007 13:30
On Wednesday 21 November 2007, Ed O'Connor wrote:
> messaging language. I think I'd be struck down to hear someone say,
> "Boy, you need to get your hands on a good, solid messaging language!"
Bam!
I can tell you - but I can't tell you where - that there are places where
rebol is used extensive for "messaging" as in TCP/IP. There's may be
more than what I know of - but there are organizations that (and big
ones at that) that may be using rebol, but are not going to advertise
it like google might advertise its use of python.
> I think that REBOL's strong point is it's convenience. I like the fact
> that I can fire up the interpreter on Windows or OS X and use it to
> manage my local data files with minimal fuss. Or that I can parse &
> extract content from web pages quickly
I use rebol extensively for web scripting not so much for for network
stuff, for that it runs rings around python. but read on....
> The areas where I find REBOL is not so simple is in managing errors,
> networking, encryption, xml and building DSLs. I'm not saying the
> power isn't there, just that you need to have a good deal of expertise
> in these areas to leverage these features.
Oh if rebol could just report the file and line number of an error!
Python has really got it down for error handling!
To elaborate on what Alessandro has said about protocols - absolutely.
And they don't have to be native, could be mezzanine or in libraries.
Speaking of libraries, what many, many who examine rebol and remark
to me is "Why is there no libraries? What is up with that? Is this really a
serious programming language?"
Given rebol.org
as a starting point, there should be libraries. Libraries that are reviewed
vetted and blessed by RT. In my opinion, rebol is crippled without that,
because fellow programmers are telling me that is one thing that rebol
is missing..
I know, I've said this before, many times on this ML.
I'll probably say it again.
tim
[12/15] from: Izkata:Comcast at: 21-Nov-2007 12:09
On Wed, 2007-11-21 at 21:54 +1300, Carl Read wrote:
> On Wednesday, 21-Novenber-2007 at 9:00:33 Petr Krenzelok wrote,
> >
<<quoted lines omitted: 18>>
> An online desktop then, but with no need of a central server owned by others?
> A Google for "online desktop" turns up plenty of attempts, but I suspect all of them
are based on central-servers. In other words, designed for lock-in. (No time to test
them. I looked at the offerings a year or so back, and some looked promising in that
they were relatively responsive, while others were horribly slow.)
I've tried 3 or 4 of them out of curiosity, and all the ones I tried
extremely laggy. I don't see how it's possible to get anything done on
them.
[13/15] from: btiffin::rogers::com at: 21-Nov-2007 23:18
Tim;
Discussions on this very issue are in progress. The user.r topic of interest
right now is Programming in the Large. That for me includes libraries;
trusted libraries. I hope to have my head around a secure Web of Trust
sequence soon, (the whole OpenPGP thing) and getting it started if it's not
too much grief. That will give us rebols a digital signature and a bank of
keys to verify such.
It is a little bit of a catch-22. So many features that are visible libraries
for most scripters are a one or two-liner built-in for REBOL. But there are
libraries we do need and those should be built and then propagated in a well
defined and well advertised way. All planned...now action? Hmmm,
action. :)
I was trying not to mention REBOL/3 in this, but even R3 will not be the
panacea for library management. That will take the community and
some "follow the leader" imho.
Cheers,
Brian
On Wednesday 21 November 2007 17:30, Tim Johnson wrote:
[14/15] from: moliad::gmail::com at: 22-Nov-2007 1:07
On Nov 21, 2007 3:53 PM, Ed O'Connor <edoconnor-gmail.com> wrote:
> Interesting discussion. I like the notion of PHP finding its niche
> with serving web pages.
<<quoted lines omitted: 3>>
> Or, likewise, "Where can I find a dynamic, highly-reflective
> metaprogramming language when I need one." :^)
REBOL is the most productive programming environment I have used for quick
and dirty stuff. yet for all its qualities, its the least documented
framework I've ever used. sometimes you realise that its as huge as .net
and you never knew!
we all know R3 will address a lot of this. and its not just talk, the guys
handling the task, really are writting down a lot more than I can remember
in prior iterations.
> I think that REBOL's strong point is it's convenience.
yep. no other language allows you to define DSL, parse strings, create
guis, define stylesheets, embed a rich client all using ONE single datatype
rich language.
I wonder why many of the serious REBOLers are not using REBOL more. So far
I have used REBOL in all environments I can think of, from async concurrent
multi -port tcp enterprise servers, XML db use, web-service clients,
web-service servers, advanced node-based graphical pipelines, neural nets
(one done entirely in REBOL :-), http clients accessing servers providing
non browser guis, I've even made a crawler robot for a site built which does
nasty anti-robot stuff (using cookies and timer + counter based redirected
urls).
I've basically built an asp contender with remark ( but much leaner and
vastly less sprawled, one very simple object model handles everything, from
resources to master templates, to db, styles, etc).
I've programmed one of the most advanced procedural node engines I've used
so far, in any platform and a lot of other things, and well, its always
easier *overall* with rebol.
its not always faster in execution, but there is soooo much less clutter in
the way of having to program against the language or its framework.
I really find that its always simpler, even when you must roll up your
sleeves to solve a little issue which REBOL doesn't support natively.
> REBOL needs to enable the type of programming experience and end-user
> apps that people are not getting from current scripting languages and
> tool-sets.
I agree, but I find it already does, to a large extent. much more than many
other systems without needing to understand much, I find. just the use of
something like save and load is a god send you quickly long for in all other
environments.
for many, rebol goes against the long entrenched idea that complex things
are more capable. some things are much less necessary in REBOL. so its
hard to accept to release some coding habits in order to adopt another
coding style.
Having done REBOLing almost exclusively for so long now, I'm finding it hard
to adapt to other languages now. simply cause everything just seems bloated
and over-engineered, when in the end, it adds nothing to the actual
delivered application.
the issue is that simple things like an open field of grass, are harder to
use in order to arrive somewhere, cause there is no obvious destination,
there is more burden on the capacity of the architect. in complex systems,
the architect is so tethered, that in fact, his role is reduced to
implementing a predefinied path.
so it is easier, but in the end, you are just as equal as every other person
using a specific system. less capable engineers/architects cannot fathom
using REBOL (I am talking experience here, hehe), cause its too open, not
structured enough...and they are actually reminded that they lack something
in what should be their primary qualities.
just like bad managers often lack the qualities of being able to trust their
own gut feelings, changing stuff, meddling and many other project/team
destructing habits.
I'm not saying R2 is perfect. some things are lacking, only that many
people downsize it, when I have found that with minimal actual
investement... its actually already capable of just about everything people
keep saying it isn't.
(this is a general reply... and not specific to your own comments... but its
a recurring thread which I just wanted to address after having done REBOL
professionaly for about 1 1/2 years almost full-time).
there are big servers running mission critical stuff, for surprising
companies... using REBOL... its just not something people advertise. they
really don't care... as long as it works.
the only issue I personnaly have is that the view engine (not vid which is
built over it and is very flexible) is a bit too closed up, because eons
ago, it was limited so that it would be the same on about 15 different
platforms. that vision has changed a bit now, so R3 will be much more open
to being adapted to the host, in order to benefit from each host's
advantages... just like it should be.
end users don't care how things are coded... but they really care when you
change the way they are used to using software on their platform.... a
common recurring comment... "what, no icon dropping? wtf?"
-MAx
[15/15] from: tim-johnsons:web at: 22-Nov-2007 11:08
On Wednesday 21 November 2007, Brian Tiffin wrote:
Hi Brian:
:-) I'm deleting all the '=20's - don't know where they come from
but very likely, the list-serv
> It is a little bit of a catch-22. So many features that are visible
> libraries for most scripters are a one or two-liner built-in for REBOL
Even the one-liners can extend the basic functionality of the mezzanine
and be very enlightening to new and( in my case anyway) old programmers.
> But there are libraries we do need and those should be built and then
Much thought to give about how to organize them, for sure.
> propagated in a well defined and well advertised way. All planned...now > >
> action? Hmmm, action. :)
You mean, like follow-up and follow-thru? Hahaha! One step at a time and
you get there.
tim
Notes
- Quoted lines have been omitted from some messages.
View the message alone to see the lines that have been omitted