Icon View Thread

The following is the text of the current message along with any replies.
Messages 1 to 10 of 15 total
Thread BPL like deploy
Sun, Mar 31 2013 5:08 PMPermanent Link

Leslie

Tim,

I was wondering if deploying the EWB components (which do not change often) as a separate file would make loading  faster? Especially mobile devices could benifit from  being large part of the code already cached, maybe even compiled.

Cheers,
Leslie
Mon, Apr 1 2013 3:46 AMPermanent Link

Matthew Jones

See my thread on "Redundant code" for the detailed explanation of why the way it is
is better. Primarily a single file, optimised for what is used, is faster for the
devices.

/Matthew Jones/
Mon, Apr 1 2013 2:27 PMPermanent Link

Leslie

Thanks Matthew, it made the picture clearer. SmileI still would like to see/do some tests at one point to compare the actual speed  difference. Compiling the stripped code of the library could be significantly slower than having the precompiled version ready in cache. The best approach I can think of would be a mixed one: if  there is a way to check the presence  of the full version of  the library a loader can load either just the app code or the one file version with the stripped library. Once it is done it  can start downloading the full version in the background as a preparation for the next time. This way we could probably have the advantages of both ways with no performance penalty at all.

Cheers,
Leslie
Mon, Apr 1 2013 2:34 PMPermanent Link

Leslie

Maybe the stripped version of the library can be executed somewhat faster then the full version. If so there can be some performance penalty for using the full version of the lib in the case when all files are fully cached. I would guess this is not much, but only profiling can tell it for sure.  

Cheers,
Leslie
Mon, Apr 1 2013 2:47 PMPermanent Link

Leslie

This is something I have googled  and  could very well be unrelated, but I do not have the time now to evaluate its relevance for the subject:

http://lostechies.com/derickbailey/2012/04/10/javascript-performance-pre-compiling-and-caching-html-templates/

Cheers,
Leslie
Mon, Apr 1 2013 2:55 PMPermanent Link

Raul

Globestar Systems

Team Elevate Team Elevate

Leslie,

Considering the current small size of the EWB compressed JS file i'm not
convinced it's even worth the effort to measure yet.

I have 2 app deployed and the compressed js is less than 500K in both
cases. This is lot less than any average web page i visit (even mobile
ones).  Assuming server has compression enabled also the actual data
sent would be a lot less.

Deployment size is always something to think about but looking at real
world numbers at the moment i don't see a need for this yet. If EWB
keeps growing and apps gets larger then maybe but otherwise this seems
like a case of premature optimization.

Raul




On 4/1/2013 2:27 PM, Leslie wrote:
> Thanks Matthew, it made the picture clearer. SmileI still would like to see/do some tests at one point to compare the actual speed  difference. Compiling the stripped code of the library could be significantly slower than having the precompiled version ready in cache. The best approach I can think of would be a mixed one: if  there is a way to check the presence  of the full version of  the library a loader can load either just the app code or the one file version with the stripped library. Once it is done it  can start downloading the full version in the background as a preparation for the next time. This way we could probably have the advantages of both ways with no performance penalty at all.
Mon, Apr 1 2013 4:52 PMPermanent Link

Tim Young [Elevate Software]

Elevate Software, Inc.

Avatar

Email timyoung@elevatesoft.com

Leslie,

<< Thanks Matthew, it made the picture clearer. SmileI still would like to
see/do some tests at one point to compare the actual speed  difference.
Compiling the stripped code of the library could be significantly slower
than having the precompiled version ready in cache. The best approach I can
think of would be a mixed one: if  there is a way to check the presence  of
the full version of  the library a loader can load either just the app code
or the one file version with the stripped library. Once it is done it  can
start downloading the full version in the background as a preparation for
the next time. This way we could probably have the advantages of both ways
with no performance penalty at all. >>

What you're missing is that the application JS is already compressed *and*
cached (after the first load).  So, unless your application code is
constantly changing, there really isn't much of a benefit to having the
runtime library split out since the application will already be cached for
the user in the same way that you expect the runtime library to be.

Tim Young
Elevate Software
www.elevatesoft.com
Mon, Apr 1 2013 5:00 PMPermanent Link

Tim Young [Elevate Software]

Elevate Software, Inc.

Avatar

Email timyoung@elevatesoft.com

Leslie,

<< This is something I have googled  and  could very well be unrelated, but
I do not have the time now to evaluate its relevance for the subject: >>

It is unrelated.  They're talking about pre-compiling JS/HTML code for use
as templates.

What you want is this discussion:

http://stackoverflow.com/questions/1096907/do-browsers-parse-javascript-on-every-page-load

and this:

http://stackoverflow.com/questions/889195/do-browsers-compile-and-cache-javascript

Tim Young
Elevate Software
www.elevatesoft.com
Mon, Apr 1 2013 5:40 PMPermanent Link

Leslie

Tim,

what I would like to speed up is exately the first time loading/reloading.  My first project with EWB was a webshop, the first time experience matters there. I am sure there is no performance issue with the PC's of late, but the target should be the low end android handsets and tablets. And that's where optimization does matter. I understand that EWB has not been optimized for mobiles yet, but I always do the user experience tests on the weakest targeted hardware so I have checked. It was simply too slow. It got me thinking  about optimizing load speed. Till that  point the speed was mostly satisfying. The speed of the development is especially great, I really enjoy developing with EWB, Smile

Cheers,
Leslie
Mon, Apr 1 2013 5:45 PMPermanent Link

Leslie

Raul,

you may be right but I still would like to see the numbers. Smile
Have you tested your apps with  low end mobiles or tablets?

Cheers,
Leslie
Page 1 of 2Next Page
Jump to Page:  1 2
Image