State of the art: Server-side JavaScript

[Alexandre Morgaut]

Hi,

I said recently that I was working on such a topic. We did a lightning
talk (15mn) version of it at ParisJS so you can have already a look to
the slides if you are interested (even if lot of the given
informations were not written).

http://www.slideshare.net/alexandre_morgaut/state-of-the-art-serverside-javascript-parisjs

I'd be very happy to have your feedbacks. Any suggestion / corrections
would be welcome.

The full version will be presented at Wakanday / JS.everywhere(Boston,
October, 15)

I might be able to have some good price tickets for attendees coming
from this group.
If anyone would be interested to sponsor it, feel free to contact me.

Regards,

Alexandre

[Wes Garland]

Alexandre;

I would like to call your attention to http://ape-project.org/ -- he deserves a mention not only because he's been around for years and has users, but because his SSJS approach/framework is quite novel - at its heart, it's a SS/Browser object/event exchange framework + "Comet" web server.

As for "JS Everywhere" - that's the entire point of the GPSEE project (and exactly how we use in real-world every day work). In my company, we have JS in web servers, mail servers, we write JS "shell scripts" for system administration batch jobs, reports, etc.  We're trying to put JS where PERL was 15 years ago.

Wes

--
You received this message because you are subscribed to the Google Groups "CommonJS" group.
To post to this group, send email to comm...@googlegroups.com.
To unsubscribe from this group, send email to commonjs+u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/commonjs?hl=en.

--
Wesley W. Garland
Director, Product Development
PageMail, Inc.
+1 613 542 2787 x 102

[Alexandre Morgaut]

Hi Wesley,

thank you for the suggestions, APE and GPSEE will be in the next
version.

[Alexandre Morgaut]

Hi,

I presented an updated version of this talk at the WebWorkerCamp in
the Open World Forum this week-end: http://www.openworldforum.org/Articles/Barcamp

The new slides are available there:
http://www.slideshare.net/alexandre_morgaut/state-of-the-art-server-side-java-script-webworkerscamp-9406069

I figured during the Q&A that I'll have to add another chapter giving
use cases like:
- Web app
- Web site
- embedded app
- faceless interface test
- proxy

Someone from "Linux Mag" said they could be interested to have en
article from it, so if any of you could tell me which solution they
use in which context, I'd be very happy to include it.

Alexandre.

[Wes Garland]

Alexandre:

• Why does the "History" slide not include Netscape Enterprise Server's Server-Side JavaScript product?
• That Husted/Kuslich book, do you know where I can buy a copy?
• Somebody in your audience might pipe and say, "SpiderMonkey is C++ now". You should answer, "Yes, but the API is still C". Similarly, "TraceMonkey", "IonMonkey", and "JaegerMonkey" might come up; these are specific milestones in SpiderMonkey development, and not separate products.
  
• In the speed comparisons stuff, it might be worthwhile to note the tremendous speed increases in JS over the years. I prepared a chart when Firefox 4 came out (which I hereby release into the public domain without restriction) that shows the trend in SpiderMonkey's SunSpider scores. http://www.page.ca/~wes/SpiderMonkey/Perf/sunspider_history.png
  

Great work! It looks like it will be a fun presentation to watch!

Wes

[Alexandre Morgaut]

>    - Why does the "History" slide not include Netscape Enterprise Server's
>    Server-Side JavaScript product?

Netscape Enterprise Server is included in the "History" section (slide
3)

>    - That Husted/Kuslich book, do you know where I can buy a copy?

I bought it via Amazon:
-
http://www.amazon.com/Server-Side-JavaScript-Developing-Integrated-Applications/dp/020143329X/ref=sr_1_1?s=books&ie=UTF8&qid=1317067785&sr=1-1

I also have the Netscape official book:
- http://www.amazon.com/Official-Netscape-Server-Side-JavaScript-Applications/dp/1566047455/ref=sr_1_3?s=books&ie=UTF8&qid=1317067964&sr=1-3


>    - Somebody in your audience might pipe and say, "SpiderMonkey is C++

>    now". You should answer, "Yes, but the API is still C". Similarly,
>    "TraceMonkey", "IonMonkey", and "JaegerMonkey" might come up; these are
>    specific milestones in SpiderMonkey development, and not separate products.

Thank you for these precisions

>    - In the speed comparisons stuff, it might be worthwhile to note the

>    tremendous speed increases in JS over the years. I prepared a chart when
>    Firefox 4 came out (which I hereby release into the public domain without
>    restriction) that shows the trend in SpiderMonkey's SunSpider scores.
>    http://www.page.ca/~wes/SpiderMonkey/Perf/sunspider_history.png

Awesome!
I include it for the next version, and hopefully for the article.

>
> Great work! It looks like it will be a fun presentation to watch!

Thank you, the public looked very excited, I think both ParisJS and
WebWorkersCamp communities did appreciate it.
The first version of the slides (for ParisJS) stayed 4/5 days on the
slideshare homepage and had more than 3,000 views in 2 weeks
The second version reached 640 in 2 days including sunday...
I'm sure some Gurus would made these scores quite bigger, but having
such numbers by myself make me think that people is more and more
interested by the Server-Side JavaScript. It motivates me to do it
better and better.

I'd really love to make this document more community driven and give
some credits.

Alexandre.

[Wes Garland]

> I include it for the next version, and hopefully for the article.

Something I forgot - JS 1.5 is ~ Firefox 2.0 (and ES-3), 1.7 is ~ Firefox 3.0, 1.8 is ~ Firefox 3.6 and 1.8.5 + JITs is ~ Firefox 4.

The benchmarks were all done with the same computer / OS / compiler / compiler settings, etc, so should be a pretty good relative measure.  I couldn't test JS 1.4 because the SunSpider benchmark requires an ES-3 capable engine.