Home » Language IDEs » ServerTools (WTP) » Web Tools Platform Proposal Has Entered the Building!
| | | | | | | |
| Re: Web Tools Platform Proposal Has Entered the Building! [message #20845 is a reply to message #20832] |
Thu, 29 April 2004 14:44   |
Eclipse User |
|
|
|
Originally posted by: ivasara-NOSPAM-.cbc.ca
It seems pretty clear that the Eclipse Webtools project won't be
touching any of the third party tools that we know and love (Spring,
Hibernate, Xdoclet, Struts, Tapestry, etc..). This leaves them to focus
on the official W3C & JCP standards which is fine by me, I'd love a
totally kickass JSP editor (myeclipse comes close... )
I have to put my support behind this seperation as it still leaves a
nice ripe market for third party plugin writers to provide tools for the
bleeding edge, nonstandard tools.
Michael Dickson wrote:
> Frankly I'd "start" based on the scope of the contributions (with the
> proposed feature areas providing definitions of how to partition and package
> the contributions) and then take a step back and see what's missing after
> that. From what I understand from the material I've seen the contributions
> from IBM are significant and given past support from readers of this group
> you'd likely get many hands to help whip everything into shape.
>
> In short something sooner is better in this case since there is a good basis
> to start from. Spending a bunch of time apple polishing to produce the
> perfect solution doesn't make sense to me. And the last thing I want to see
> is a whole new set of "design-ed frameworks" when a good existing starting
> point exists.
>
> Mike
>
> "Christophe Ney" <christophe.ney@objectweb.org> wrote in message
> news:c6ohoa$dr0$1@eclipse.org...
>
>>Hi Haris,
>>
>>I understand that you want more, but the scope of Web tooling is huge and
>
> we
>
>>have choosen to propose to start with something for which the scope is
>
> very
>
>>well
>>identified.
>>
>>That said, we really want the result of this project to be extensible and
>
> to
>
>>ease the
>>development of tools such as the one you describe. If you have interest in
>>those,
>>tools you should get involved and help us to design frameworks and APIs
>
> that
>
>>work for this.
>>
>>Sincerely,
>>Christophe
>>
>>"snpe" <snpe@snpe.co.yu> a écrit dans le message news:
>>c6mntd$cth$1@eclipse.org...
>>
>>>My wish list :
>>>JSF (included)
>>>hibernate
>>>spring
>>>
>>>regards
>>>Haris Peco
>>
>>
>>
>
>
|
|
|
| Re: Web Tools Platform Proposal Has Entered the Building! [message #21198 is a reply to message #20640] |
Wed, 05 May 2004 11:08 |
| Eclipse User |
|
|
|
Congratulations,
I'm delighted, that the separation of web-standard supporting tools and
J2EE tools has prevailed in the discussions and the overall design. This
opens this project to much broader use, thanif it would have been
started with the J2EE focus, just because it is there allready.
A few observations:
- I'm missing among the web-artifacts the graphical artifacts, such as
*.png, *.gif, etc.. They might be concidered as important and need some
support in form of editors.
- It might be worth mentioning other web-artifacts in genral such as
cgi-bin scripts, robots.txt, .htaccess, etc. There could be a whole
cottage industry to build editors for those and I believe it would be
appreciated to have them added to the project in general. May be even
the server modules should throw warning, if one deployes artifacts that
are not recogniced there (such as .htaccess to a Netscape server).
- Another set of popular formats such as *.pdf and *.svg could be added
to. Although they are not all web standards in the W3C sense.
- I'm missing also some higher notion of page artifact such as a
template (static and dynamic) and some two stage content generation
process (merge template with content - may be from multiple sources -
and generate the actual HTML code) so the results of dynamic structures
can be viewed and tested.
- I wonder how Cocoon relates to J2EE? I thought it is totally
independend of it. And isn't Struts as well independent of J2EE? I guess
its just because they are used so much on top of J2EE.
- Can someone explain what the use case "Deploy an applet on a given
HTTP server" means? In my understanding Java applets are server
independent, aren't they? Or do we talk a different type of applet
- I'd like to see a section added to discuss the refactoring framework
in the standard web tools project. Because it is not trivial and there
are wise things to do, such as never delete any deployed (versioned)
URL, but replace the page with a redirect instead. Sure one could do
that on the meta level, but refactoring support for such things would be
really a great help. May be there could be some integrety rules, that
are customizable, so one could run a tighter ship if one desires.
- Is there an idea of a generic html service infrastructure around. Web
services are great but there are so many tools available on the net,
from validating HTML (probably a rather build in function) to meassuring
keyword density or search engine positions, that could be exploited by
integration. Having an API level to push dialogs via HTML would help to
make these things available within the IDE.
- Last but not least, I'm missing any testing models. Do we hope they
fall in place or do we plan to integrate some tooling around that? How
about test data artifacts, etc.?
Thanks for listening to my arbitrary comments. I look forward to start
using some of the tools and help out in testing.
Kaj
Arthur Ryman wrote:
> The proposal has been posted at
> http://www.eclipse.org/proposals/eclipse-webtools/index.html
>
> I'd personally like to thank Bjorn Freeman-Benson and John Wiegand for
> guiding us through the new Eclipse Development Process and getting the
> project activated.
>
> I'd also like to thank Christophe Ney for pulling the proposal and the
> participants together.
>
> -- Arthur
>
|
|
|
| Re: Web Tools Platform Proposal Has Entered the Building! [message #21540 is a reply to message #21198] |
Sun, 09 May 2004 08:37 |
| Eclipse User |
|
|
|
Kay,
Thx for the feedback. My responses are below:
exquisitus wrote:
> Congratulations,
> I'm delighted, that the separation of web-standard supporting tools and
> J2EE tools has prevailed in the discussions and the overall design. This
> opens this project to much broader use, thanif it would have been
> started with the J2EE focus, just because it is there allready.
>
> A few observations:
> - I'm missing among the web-artifacts the graphical artifacts, such as
> *.png, *.gif, etc.. They might be concidered as important and need some
> support in form of editors.
I think image tools are in scope although we don't have any
contributions on the table. I think we should give preference to PNG
since it is a W3C standard.
> - It might be worth mentioning other web-artifacts in genral such as
> cgi-bin scripts, robots.txt, .htaccess, etc. There could be a whole
> cottage industry to build editors for those and I believe it would be
> appreciated to have them added to the project in general. May be even
> the server modules should throw warning, if one deployes artifacts that
> are not recogniced there (such as .htaccess to a Netscape server).
Agreed. I think tools to support Open Source Web servers are in scope.
And of course we should make these extensions of some core tools so
other vendors are on an equal footing. Part of the proposed IBM
contribution is a Structure Source Editor (SSE) which might be
applicable to the source formats used for server config.
> - Another set of popular formats such as *.pdf and *.svg could be added
> to. Although they are not all web standards in the W3C sense.
SVG is a W3C specification and SVG tools would be in scope although we
have no contribution on the table. PDF is an Adobe technology, but a
defacto industry standard. A nicely integrated PDF viewer would be good.
Eclipse does let you launch Adobe Acrobat, but it is weakly integrated.
> - I'm missing also some higher notion of page artifact such as a
> template (static and dynamic) and some two stage content generation
> process (merge template with content - may be from multiple sources -
> and generate the actual HTML code) so the results of dynamic structures
> can be viewed and tested.
My assumption here is that we would enable XSTL tools, which is the W3C
standard way to generate static content, and JSP and JSF which are the
J2EE ways to do dynamic content. We would enable other template
technologies, like Velocity.
> - I wonder how Cocoon relates to J2EE? I thought it is totally
> independend of it. And isn't Struts as well independent of J2EE? I guess
> its just because they are used so much on top of J2EE.
Struts depends on many J2EE standards, e.g. servlets, JSP, tag libs.
Cocoon is based on servlets and XML. So they both would be enabled by a
level of standards based, extensible tools.
> - Can someone explain what the use case "Deploy an applet on a given
> HTTP server" means? In my understanding Java applets are server
> independent, aren't they? Or do we talk a different type of applet
Yes, applets are client side and part of J2SE. However, you still need
to generate an HTML page with <applet> or <object> tags, and it would be
nice if Eclipse could attach its debugger to the JVM running inside IE
or Mozilla rather than just the AppletViewer. At present, Eclipse has no
concept of "Web Application". The Web Standard tools should extend the
current JDT applet development support into the context of Web servers
and browsers.
> - I'd like to see a section added to discuss the refactoring framework
> in the standard web tools project. Because it is not trivial and there
> are wise things to do, such as never delete any deployed (versioned)
> URL, but replace the page with a redirect instead. Sure one could do
> that on the meta level, but refactoring support for such things would be
> really a great help. May be there could be some integrety rules, that
> are customizable, so one could run a tighter ship if one desires.
Good thoughts. Refactoring is very open ended. Defining the key
operations in the context of Web apps is new ground. Current tools do
things like updating URLs. A comprehensive list of operations would be
very useful. We need a Martin Fowler for the Web App domain.
> - Is there an idea of a generic html service infrastructure around. Web
> services are great but there are so many tools available on the net,
> from validating HTML (probably a rather build in function) to meassuring
> keyword density or search engine positions, that could be exploited by
> integration. Having an API level to push dialogs via HTML would help to
> make these things available within the IDE.
HTML validation is certainly in scope. On you point about HTML dialogs,
do you mean generating Eclipse UI via HTML? If so, there is a mechanism
thet IBM exploits for some tools. The Help system uses an integrated
servlet engine (Tomcat) to display Help. IBM uses it to render the Web
Service Explorer. The UI is generated by JSP and is integrated with the
Eclipse platform, e.g. I can import a WSDL into my project.
> - Last but not least, I'm missing any testing models. Do we hope they
> fall in place or do we plan to integrate some tooling around that? How
> about test data artifacts, etc.?
Testing in general is the domain of the Eclipse Hyades project. However,
there are some specific extensions that belong in WTP, e.g. WS-I
compliance checking for Web services. I believe Hyades has some HTTP
test tools, but certainly WTP could include some good integration for
launching the Hyades test tools.
>
> Thanks for listening to my arbitrary comments. I look forward to start
> using some of the tools and help out in testing.
>
> Kaj
-- Arthur
|
|
|
| Re: Web Tools Platform Proposal Has Entered the Building! [message #21771 is a reply to message #21540] |
Mon, 10 May 2004 10:43 |
| Eclipse User |
|
|
|
Hi Arthur,
thanks for your very encouraging comments. I'll add a few thoughts and
answers to your questions below:
Arthur Ryman wrote:
> Kay,
>
> Thx for the feedback. My responses are below:
>
> exquisitus wrote:
>...
>> - It might be worth mentioning other web-artifacts in genral such as
>> cgi-bin scripts, robots.txt, .htaccess, etc. There could be a whole
>> cottage industry to build editors for those and I believe it would be
>> appreciated to have them added to the project in general. May be even
>> the server modules should throw warning, if one deployes artifacts
>> that are not recogniced there (such as .htaccess to a Netscape server).
>
>
> Agreed. I think tools to support Open Source Web servers are in scope.
> And of course we should make these extensions of some core tools so
> other vendors are on an equal footing. Part of the proposed IBM
> contribution is a Structure Source Editor (SSE) which might be
> applicable to the source formats used for server config.
>
I look forward to see what SSE is covering. I guess of great imnortance
here would be good tutorials so volunteers can use such tools
efficiently and create as many descriptions of file formats (and
specialized editors) as possible.
>> - Another set of popular formats such as *.pdf and *.svg could be added
>> to. Although they are not all web standards in the W3C sense.
>
>
> SVG is a W3C specification and SVG tools would be in scope although we
> have no contribution on the table. PDF is an Adobe technology, but a
> defacto industry standard. A nicely integrated PDF viewer would be good.
> Eclipse does let you launch Adobe Acrobat, but it is weakly integrated.
May be you can get Adobe in the boat to support Eclipse with a tighter
integration?
>> - I'm missing also some higher notion of page artifact such as a
>> template (static and dynamic) and some two stage content generation
>> process (merge template with content - may be from multiple sources -
>> and generate the actual HTML code) so the results of dynamic
>> structures can be viewed and tested.
>
>
> My assumption here is that we would enable XSTL tools, which is the W3C
> standard way to generate static content, and JSP and JSF which are the
> J2EE ways to do dynamic content. We would enable other template
> technologies, like Velocity.
>
As there are so many systems generating pages from templates, starting
with copying a template as a starting point for creating a static page,
over simple variable substitution to xml generation and CMS systems etc.
It would be good to build structure according to this into the projects,
so one can distinguish between the objects ultimately going to be
deployed and the ones that are intermediary artifacts to build the site
>> - I wonder how Cocoon relates to J2EE? I thought it is totally
>> independend of it. And isn't Struts as well independent of J2EE? I
>> guess its just because they are used so much on top of J2EE.
>
>
> Struts depends on many J2EE standards, e.g. servlets, JSP, tag libs.
> Cocoon is based on servlets and XML. So they both would be enabled by a
> level of standards based, extensible tools.
>
I would distinguish between using Cocoon, which is solely xml and the
internal sturctures of the xmaps etc. and developing additional
components for Cocoon, which off course is a servlet development task.
Just because eclipse uses xml files and xerces as its xml parser, in my
view it does not constitute a xml development system (proof in case -
that is part of this project) or is dependent on an xml development
system. In analogy a cocoon web-site development system is not dependent
on a JSP development system (--> like sunBow lives happily without it).
This changes, when one wants to develop components for Cocoon, but that
is a much smaller user community.
>> - Can someone explain what the use case "Deploy an applet on a given
>> HTTP server" means? In my understanding Java applets are server
>> independent, aren't they? Or do we talk a different type of applet
>
>
> Yes, applets are client side and part of J2SE. However, you still need
> to generate an HTML page with <applet> or <object> tags, and it would be
> nice if Eclipse could attach its debugger to the JVM running inside IE
> or Mozilla rather than just the AppletViewer. At present, Eclipse has no
> concept of "Web Application". The Web Standard tools should extend the
> current JDT applet development support into the context of Web servers
> and browsers.
>
I see, I was just confused by the reference to "a given HTTP server".
I'd think the goals could be better described in the proposal, as you
did here.
>> - I'd like to see a section added to discuss the refactoring framework
>> in the standard web tools project. Because it is not trivial and there
>> are wise things to do, such as never delete any deployed (versioned)
>> URL, but replace the page with a redirect instead. Sure one could do
>> that on the meta level, but refactoring support for such things would
>> be really a great help. May be there could be some integrety rules,
>> that are customizable, so one could run a tighter ship if one desires.
>
>
> Good thoughts. Refactoring is very open ended. Defining the key
> operations in the context of Web apps is new ground. Current tools do
> things like updating URLs. A comprehensive list of operations would be
> very useful. We need a Martin Fowler for the Web App domain.
>
>> - Is there an idea of a generic html service infrastructure around.
>> Web services are great but there are so many tools available on the
>> net, from validating HTML (probably a rather build in function) to
>> meassuring keyword density or search engine positions, that could be
>> exploited by integration. Having an API level to push dialogs via HTML
>> would help to make these things available within the IDE.
>
>
> HTML validation is certainly in scope. On you point about HTML dialogs,
> do you mean generating Eclipse UI via HTML? If so, there is a mechanism
> thet IBM exploits for some tools. The Help system uses an integrated
> servlet engine (Tomcat) to display Help. IBM uses it to render the Web
> Service Explorer. The UI is generated by JSP and is integrated with the
> Eclipse platform, e.g. I can import a WSDL into my project.
>
My thoughts were more in the direction of autocompleting forms and
automating dialogs, so one can easily use html based services. For
example, given a HTML page (pulled from a web-server) fill in the form
values from the context of a project (like the HTML code of a page for
meassuring keyword density). Then display the resulting page in a
browser or even store the resulting page or extract parts of it as
historical data.
One could off course simply try to make all these services accessible by
webservices as well. Unfortunately such webservices are rare and i
twill tak ea recolution to get there. I see a great knowledge barrier to
generate such web services and there is also a hosting barrier. After
all web hosting costs less than $10 a month (affordable for a hobbyist),
but hosting any web services app is much more expensive, despite the
fact that resources required should be less in most cases (eliminate all
the HTML overhead for rendering).
May be I'm a bit ambitious here. I just see the vast capabilities of
such tools on the Internet and they are inaccessible on a project basis.
If I want to check one web page for keyword density (just my favorite
example), the usual dialogs are fine. However, if I have to check 50
pages, I'd like to do this with a review (like post build) of my
project, automated, and listing warning messages in Eclipse. Can you
share my dream here?
Reviewing this topic brings me to the question: Will web services
integration be a part of Eclipse? So far I have seen only the idea of
developing web-services. Many functions could be gained in a development
environment, by adding access to web-services in order to access remote
functions. For example one could access spell checkers, thesaurus, check
search engine positions for keywords and URLs, etc. Of great help would
be the ability to define a request with artifacts from the IDE and to
interprete the results by defining trigger mechanisms to display results
in search lists, task lists or as markers in editors.
>> - Last but not least, I'm missing any testing models. Do we hope they
>> fall in place or do we plan to integrate some tooling around that? How
>> about test data artifacts, etc.?
>
>
> Testing in general is the domain of the Eclipse Hyades project. However,
> there are some specific extensions that belong in WTP, e.g. WS-I
> compliance checking for Web services. I believe Hyades has some HTTP
> test tools, but certainly WTP could include some good integration for
> launching the Hyades test tools.
>
I'll get a look at Hyades!
>>
>> Thanks for listening to my arbitrary comments. I look forward to start
>> using some of the tools and help out in testing.
>>
>> Kaj
>
>
> -- Arthur
>
I can't wait to start looking at the first code base and contributing a
little of my testing time.
Kaj
|
|
|
Goto Forum:
Current Time: Wed Jul 02 22:47:59 EDT 2025
Powered by FUDForum. Page generated in 0.05545 seconds |
Home