Home » Eclipse Projects » Remote Application Platform (RAP) » GWT vs RAP and some random ideas [Importance: Low]
|
Re: GWT vs RAP and some random ideas [Importance: Low] [message #47915 is a reply to message #47758] |
Wed, 19 September 2007 10:29 |
Benjamin Muskalla Messages: 237 Registered: July 2009 |
Senior Member |
|
|
Hi Joel,
see my comments below.
Joel Oliveira wrote:
> As stated before in this newsgroup, almost everyone that is evaluating
> RAP, also knows about GWT. The choice between this 2 frameworks it's,
> most of the time, based on the "look and feel" and available widgets.
I don't think that these are the only criterias. I think most of the RAP
user come from a RCP background and don't want to learn another
framework for their application developement. Also the fact that you can
migrate an existing RCP application into a web application is a big plus
here. For people with this RCP background you may be right with the
"look and feel" and the available widgets. But besides the widgets RAP
has the advantage to support all concepts of JFace and the Workbench
(Viewer, Views, Perspectives, Editors, etc).
>
> In my case, even if I'm pushing RAP in one of my newest projects...
Nice to hear!
>
> I posted a question in a TSS thread
> ( http://www.theserverside.com/news/thread.tss?thread_id=46763 #239327)
> about the new GWT release vs RAP and received some performance complains
> about RAP demo
> ( http://www.theserverside.com/news/thread.tss?thread_id=46763 #239336 "I
> just tried the RAP demo, it's so slow to be unusable"). Is it related
> with the server where it's hosted? Or is it related with the RAP itself?
> I know that you have done some performance improvements for the last
> versions.
There are some aspects why it is slow. First of all it's not the best
server where the demo is hosted. It is initially only a small one for
the company website and not really an enviroment for such an
application. The other point is that the RAP version behind that demo is
really old (i think M4 or M5) as we didn't had the time yet to update
the online demo - sorry! In the last weeks we cutted down qooxdoo a
little bit to make it faster. Qooxdoo is the real problem because it has
so much functionality not needed by RAP but coupled so tightly that we
cannot remove it without having major problems with the next update. As
you can see in the org.eclipse.rap.tools project we are stripping down
qooxdoo as far as we can. We did also many stress tests and performance
measurements in the last weeks and fixed (and still fixing) some of the
server-side problems (which are basically not really that fatal.
> New gwt-ext project it's also providing the missing "nice look" into the
> GWT widgets
> ( http://gwt-ext.googlecode.com/svn/trunk/site/samples/Showcas e/www/com.gwtext.sample.showcase.Showcase/Showcase.html).
>
> Take a look into Editable Grid and Live Search as they are missing in
> the actual RAP version.
We know that these things are still missing. As we are just about to
release 1.0 of RAP I think it's ok to not support everything from the
beginnging ;) But you're right that it is needed and we are looking for
solutions. The main problems are not the implementation on the client
side but more the compatibility with existing API from SWT and JFace.
But the biggest problem is still time :)
>
> But these are just my 5 cents,
>
> Keep the good work!
Thanks for the 5 cents - we really prefer some critic instead of people
going away without telling us how to do our job better!
Greets
Benny
|
|
|
Re: GWT vs RAP and some random ideas [Importance: Low] [message #48037 is a reply to message #47915] |
Wed, 19 September 2007 11:17 |
Benjamin Muskalla Messages: 237 Registered: July 2009 |
Senior Member |
|
|
Benjamin Muskalla wrote:
> Hi Joel,
>
> see my comments below.
>
> Joel Oliveira wrote:
>> As stated before in this newsgroup, almost everyone that is evaluating
>> RAP, also knows about GWT. The choice between this 2 frameworks it's,
>> most of the time, based on the "look and feel" and available widgets.
> I don't think that these are the only criterias. I think most of the RAP
> user come from a RCP background and don't want to learn another
> framework for their application developement. Also the fact that you can
> migrate an existing RCP application into a web application is a big plus
> here. For people with this RCP background you may be right with the
> "look and feel" and the available widgets. But besides the widgets RAP
> has the advantage to support all concepts of JFace and the Workbench
> (Viewer, Views, Perspectives, Editors, etc).
>>
>> In my case, even if I'm pushing RAP in one of my newest projects...
> Nice to hear!
>>
>> I posted a question in a TSS thread
>> ( http://www.theserverside.com/news/thread.tss?thread_id=46763 #239327)
>> about the new GWT release vs RAP and received some performance
>> complains about RAP demo
>> ( http://www.theserverside.com/news/thread.tss?thread_id=46763 #239336
>> "I just tried the RAP demo, it's so slow to be unusable"). Is it
>> related with the server where it's hosted? Or is it related with the
>> RAP itself? I know that you have done some performance improvements
>> for the last versions.
> There are some aspects why it is slow. First of all it's not the best
> server where the demo is hosted. It is initially only a small one for
> the company website and not really an enviroment for such an
> application. The other point is that the RAP version behind that demo is
> really old (i think M4 or M5) as we didn't had the time yet to update
> the online demo - sorry! In the last weeks we cutted down qooxdoo a
> little bit to make it faster. Qooxdoo is the real problem because it has
> so much functionality not needed by RAP but coupled so tightly that we
> cannot remove it without having major problems with the next update. As
> you can see in the org.eclipse.rap.tools project we are stripping down
> qooxdoo as far as we can. We did also many stress tests and performance
> measurements in the last weeks and fixed (and still fixing) some of the
> server-side problems (which are basically not really that fatal.
>
>> New gwt-ext project it's also providing the missing "nice look" into
>> the GWT widgets
>> ( http://gwt-ext.googlecode.com/svn/trunk/site/samples/Showcas e/www/com.gwtext.sample.showcase.Showcase/Showcase.html).
>>
>> Take a look into Editable Grid and Live Search as they are missing in
>> the actual RAP version.
> We know that these things are still missing. As we are just about to
> release 1.0 of RAP I think it's ok to not support everything from the
> beginnging ;) But you're right that it is needed and we are looking for
> solutions. The main problems are not the implementation on the client
> side but more the compatibility with existing API from SWT and JFace.
> But the biggest problem is still time :)
Just wanted to add something regarding "nice look". You can style your
application however you want - see http://wiki.eclipse.org/RAP_Theming
>
>>
>> But these are just my 5 cents,
>>
>> Keep the good work!
> Thanks for the 5 cents - we really prefer some critic instead of people
> going away without telling us how to do our job better!
>
> Greets
> Benny
|
|
|
Re: GWT vs RAP and some random ideas [Importance: Low] [message #48127 is a reply to message #48037] |
Wed, 19 September 2007 12:02 |
Eclipse User |
|
|
|
Originally posted by: just4lists.nospammail.net
See my comments inline.
"Benjamin Muskalla" <bmuskalla@innoopract.com> escreveu na mensagem
news:fcr0lf$21s$1@build.eclipse.org...
> Benjamin Muskalla wrote:
>> Hi Joel,
>>
>> see my comments below.
>>
>> Joel Oliveira wrote:
>>> As stated before in this newsgroup, almost everyone that is evaluating
>>> RAP, also knows about GWT. The choice between this 2 frameworks it's,
>>> most of the time, based on the "look and feel" and available widgets.
>> I don't think that these are the only criterias. I think most of the RAP
>> user come from a RCP background and don't want to learn another framework
>> for their application developement. Also the fact that you can migrate an
>> existing RCP application into a web application is a big plus here. For
>> people with this RCP background you may be right with the "look and feel"
>> and the available widgets. But besides the widgets RAP has the advantage
>> to support all concepts of JFace and the Workbench (Viewer, Views,
>> Perspectives, Editors, etc).
Agree. It was only a "simplified" view on this subject. _;) I believe "we",
as RAP community, can help the RAP team also in the task of disseminate info
about the project. At least, from talks with all my colleagues (from several
companies), GWT it's from far much more known. Some times they are just
unaware of RAP. Something that I'm trying to fix. _:)
>>> In my case, even if I'm pushing RAP in one of my newest projects...
>> Nice to hear!
>>>
>>> I posted a question in a TSS thread
>>> ( http://www.theserverside.com/news/thread.tss?thread_id=46763 #239327)
>>> about the new GWT release vs RAP and received some performance complains
>>> about RAP demo
>>> ( http://www.theserverside.com/news/thread.tss?thread_id=46763 #239336 "I
>>> just tried the RAP demo, it's so slow to be unusable"). Is it related
>>> with the server where it's hosted? Or is it related with the RAP itself?
>>> I know that you have done some performance improvements for the last
>>> versions.
>> There are some aspects why it is slow. First of all it's not the best
>> server where the demo is hosted. It is initially only a small one for the
>> company website and not really an enviroment for such an application. The
>> other point is that the RAP version behind that demo is really old (i
>> think M4 or M5) as we didn't had the time yet to update the online demo -
>> sorry! In the last weeks we cutted down qooxdoo a little bit to make it
>> faster. Qooxdoo is the real problem because it has so much functionality
>> not needed by RAP but coupled so tightly that we cannot remove it without
>> having major problems with the next update. As you can see in the
>> org.eclipse.rap.tools project we are stripping down qooxdoo as far as we
>> can. We did also many stress tests and performance measurements in the
>> last weeks and fixed (and still fixing) some of the server-side problems
>> (which are basically not really that fatal.
Nice to know.
>>> New gwt-ext project it's also providing the missing "nice look" into the
>>> GWT widgets
>>> ( http://gwt-ext.googlecode.com/svn/trunk/site/samples/Showcas e/www/com.gwtext.sample.showcase.Showcase/Showcase.html).
>>> Take a look into Editable Grid and Live Search as they are missing in
>>> the actual RAP version.
>> We know that these things are still missing. As we are just about to
>> release 1.0 of RAP I think it's ok to not support everything from the
>> beginnging ;) But you're right that it is needed and we are looking for
>> solutions. The main problems are not the implementation on the client
>> side but more the compatibility with existing API from SWT and JFace. But
>> the biggest problem is still time :)
> Just wanted to add something regarding "nice look". You can style your
> application however you want - see http://wiki.eclipse.org/RAP_Theming
I know. We are doing it.
>>> But these are just my 5 cents,
>>>
>>> Keep the good work!
>> Thanks for the 5 cents - we really prefer some critic instead of people
>> going away without telling us how to do our job better!
More than just provide some input, we will also try to provide some real
patches or widgets to the community.
Regards,
Joel Oliveira
|
|
|
Re: GWT vs RAP and some random ideas [Importance: Low] [message #48180 is a reply to message #48037] |
Wed, 19 September 2007 12:13 |
Eclipse User |
|
|
|
Originally posted by: just4lists.nospammail.net
Regarding the performance issue. Do you/have any performance numbers
regarding the 2 different approachs (Standalone VS Integrated into Tomcat)?
Best regards,
Joel Oliveira
"Benjamin Muskalla" <bmuskalla@innoopract.com> escreveu na mensagem
news:fcr0lf$21s$1@build.eclipse.org...
> Benjamin Muskalla wrote:
>> Hi Joel,
>>
>> see my comments below.
>>
>> Joel Oliveira wrote:
>>> As stated before in this newsgroup, almost everyone that is evaluating
>>> RAP, also knows about GWT. The choice between this 2 frameworks it's,
>>> most of the time, based on the "look and feel" and available widgets.
>> I don't think that these are the only criterias. I think most of the RAP
>> user come from a RCP background and don't want to learn another framework
>> for their application developement. Also the fact that you can migrate an
>> existing RCP application into a web application is a big plus here. For
>> people with this RCP background you may be right with the "look and feel"
>> and the available widgets. But besides the widgets RAP has the advantage
>> to support all concepts of JFace and the Workbench (Viewer, Views,
>> Perspectives, Editors, etc).
>>>
>>> In my case, even if I'm pushing RAP in one of my newest projects...
>> Nice to hear!
>>>
>>> I posted a question in a TSS thread
>>> ( http://www.theserverside.com/news/thread.tss?thread_id=46763 #239327)
>>> about the new GWT release vs RAP and received some performance complains
>>> about RAP demo
>>> ( http://www.theserverside.com/news/thread.tss?thread_id=46763 #239336 "I
>>> just tried the RAP demo, it's so slow to be unusable"). Is it related
>>> with the server where it's hosted? Or is it related with the RAP itself?
>>> I know that you have done some performance improvements for the last
>>> versions.
>> There are some aspects why it is slow. First of all it's not the best
>> server where the demo is hosted. It is initially only a small one for the
>> company website and not really an enviroment for such an application. The
>> other point is that the RAP version behind that demo is really old (i
>> think M4 or M5) as we didn't had the time yet to update the online demo -
>> sorry! In the last weeks we cutted down qooxdoo a little bit to make it
>> faster. Qooxdoo is the real problem because it has so much functionality
>> not needed by RAP but coupled so tightly that we cannot remove it without
>> having major problems with the next update. As you can see in the
>> org.eclipse.rap.tools project we are stripping down qooxdoo as far as we
>> can. We did also many stress tests and performance measurements in the
>> last weeks and fixed (and still fixing) some of the server-side problems
>> (which are basically not really that fatal.
>>
>>> New gwt-ext project it's also providing the missing "nice look" into the
>>> GWT widgets
>>> ( http://gwt-ext.googlecode.com/svn/trunk/site/samples/Showcas e/www/com.gwtext.sample.showcase.Showcase/Showcase.html).
>>> Take a look into Editable Grid and Live Search as they are missing in
>>> the actual RAP version.
>> We know that these things are still missing. As we are just about to
>> release 1.0 of RAP I think it's ok to not support everything from the
>> beginnging ;) But you're right that it is needed and we are looking for
>> solutions. The main problems are not the implementation on the client
>> side but more the compatibility with existing API from SWT and JFace. But
>> the biggest problem is still time :)
> Just wanted to add something regarding "nice look". You can style your
> application however you want - see http://wiki.eclipse.org/RAP_Theming
>>
>>>
>>> But these are just my 5 cents,
>>>
>>> Keep the good work!
>> Thanks for the 5 cents - we really prefer some critic instead of people
>> going away without telling us how to do our job better!
>>
>> Greets
>> Benny
|
|
|
Re: GWT vs RAP and some random ideas [Importance: Low] [message #49025 is a reply to message #48180] |
Sun, 23 September 2007 08:53 |
Eclipse User |
|
|
|
Originally posted by: fappel.innoopract.com
Hi,
I just want to make some clarifications on the perfomance problems. The demo
application that is available at http://rap.innoopract.com/rapdemo/rap is
using the M5 milestone release. With M5 we reached a development stage were
we had added most of the features we wanted to provide for the final
release. But there were no optimizations to performance and still there are
many little anoying bugs in it.
The main reason, why we did not replace the demo is because it's outdated
and would need maintainance and overwork. In particular as we want to
provide a new Demo with the upcomming 1.0 release we decided to live with
the old one for the time being.
As I am currently staying in Canada and the demo is hosted in Germany, I
took a brief look at the old demo again. The main performance problem is the
long startup time. We did a huge improvement on this with M6 by
concatenating, zipping and caching the javascript libraries and we will
spent some further work on this the comming week.
The last week we spent a lot of time on request performance optimization and
the stress-test results we got so far are very good. As our 1.0 release has
unfortunately been delayed for formality reasons, we take the chance to use
the additional time to confirm those results.
Ciao
Frank
"Joel Oliveira" <just4lists@nospammail.net> schrieb im Newsbeitrag
news:fcr3qe$fra$1@build.eclipse.org...
> Regarding the performance issue. Do you/have any performance numbers
> regarding the 2 different approachs (Standalone VS Integrated into
> Tomcat)?
>
> Best regards,
>
> Joel Oliveira
>
> "Benjamin Muskalla" <bmuskalla@innoopract.com> escreveu na mensagem
> news:fcr0lf$21s$1@build.eclipse.org...
>> Benjamin Muskalla wrote:
>>> Hi Joel,
>>>
>>> see my comments below.
>>>
>>> Joel Oliveira wrote:
>>>> As stated before in this newsgroup, almost everyone that is evaluating
>>>> RAP, also knows about GWT. The choice between this 2 frameworks it's,
>>>> most of the time, based on the "look and feel" and available widgets.
>>> I don't think that these are the only criterias. I think most of the RAP
>>> user come from a RCP background and don't want to learn another
>>> framework for their application developement. Also the fact that you can
>>> migrate an existing RCP application into a web application is a big plus
>>> here. For people with this RCP background you may be right with the
>>> "look and feel" and the available widgets. But besides the widgets RAP
>>> has the advantage to support all concepts of JFace and the Workbench
>>> (Viewer, Views, Perspectives, Editors, etc).
>>>>
>>>> In my case, even if I'm pushing RAP in one of my newest projects...
>>> Nice to hear!
>>>>
>>>> I posted a question in a TSS thread
>>>> ( http://www.theserverside.com/news/thread.tss?thread_id=46763 #239327)
>>>> about the new GWT release vs RAP and received some performance
>>>> complains about RAP demo
>>>> ( http://www.theserverside.com/news/thread.tss?thread_id=46763 #239336 "I
>>>> just tried the RAP demo, it's so slow to be unusable"). Is it related
>>>> with the server where it's hosted? Or is it related with the RAP
>>>> itself? I know that you have done some performance improvements for the
>>>> last versions.
>>> There are some aspects why it is slow. First of all it's not the best
>>> server where the demo is hosted. It is initially only a small one for
>>> the company website and not really an enviroment for such an
>>> application. The other point is that the RAP version behind that demo is
>>> really old (i think M4 or M5) as we didn't had the time yet to update
>>> the online demo - sorry! In the last weeks we cutted down qooxdoo a
>>> little bit to make it faster. Qooxdoo is the real problem because it has
>>> so much functionality not needed by RAP but coupled so tightly that we
>>> cannot remove it without having major problems with the next update. As
>>> you can see in the org.eclipse.rap.tools project we are stripping down
>>> qooxdoo as far as we can. We did also many stress tests and performance
>>> measurements in the last weeks and fixed (and still fixing) some of the
>>> server-side problems (which are basically not really that fatal.
>>>
>>>> New gwt-ext project it's also providing the missing "nice look" into
>>>> the GWT widgets
>>>> ( http://gwt-ext.googlecode.com/svn/trunk/site/samples/Showcas e/www/com.gwtext.sample.showcase.Showcase/Showcase.html).
>>>> Take a look into Editable Grid and Live Search as they are missing in
>>>> the actual RAP version.
>>> We know that these things are still missing. As we are just about to
>>> release 1.0 of RAP I think it's ok to not support everything from the
>>> beginnging ;) But you're right that it is needed and we are looking for
>>> solutions. The main problems are not the implementation on the client
>>> side but more the compatibility with existing API from SWT and JFace.
>>> But the biggest problem is still time :)
>> Just wanted to add something regarding "nice look". You can style your
>> application however you want - see http://wiki.eclipse.org/RAP_Theming
>>>
>>>>
>>>> But these are just my 5 cents,
>>>>
>>>> Keep the good work!
>>> Thanks for the 5 cents - we really prefer some critic instead of people
>>> going away without telling us how to do our job better!
>>>
>>> Greets
>>> Benny
>
|
|
| |
Re: GWT vs RAP and some random ideas [Importance: Low] [message #49154 is a reply to message #49126] |
Mon, 24 September 2007 01:22 |
Eclipse User |
|
|
|
Originally posted by: fappel.innoopract.com
Hi,
you are right, it's quite easy. In fact we will provide a small tooling set
for the upcomming 1.0 release that will contain a template to create a
"RAP-Mail" demo-app with the new project wizard,. Unfortunately this doesn't
demonstrate the distributed nature of RAP applications very well, so I don't
like it very much as a general RAP-demo - but you are right, we should
provide it at least as an additional demo. Hopefully I find some time this
week and don't forget to do it...
Ciao
Frank
"Chris Aniszczyk" <zx@us.ibm.com> schrieb im Newsbeitrag
news:fd5s6f$3s6$1@build.eclipse.org...
> May I recommend the team use the RCP Mail template as one demo? Should be
> really easy to setup. I don't like the current demo ;)
>
> Cheers,
>
> ~ Chris
|
|
|
Goto Forum:
Current Time: Sun Dec 08 00:16:59 GMT 2024
Powered by FUDForum. Page generated in 0.04063 seconds
|