Consistency is good, but I wouldn't say that's our #1 priority. We definitely want primary and secondary best practices like we do with other things, but also want to support a variety of technologies on the framework level. In other words, if people want to do stuff in their own code bases we don't want to block that, but yes for anything going into the project it should follow a consistent set of patterns so that people exploring the code for the first time will have a stream of "okay, okay, okay, ..." going through their heads instead of a stream of "wtf, wtf, wtf, ...". On an important but less pleasant note: the AJAX stuff we are dealing with is very new in OFBiz, and outside of OFBiz the "market" is still maturing quite a lot. I agree we should perhaps vote on a best practice, though in general I'd rather see X task implemented with DOJO required so much effort and with Prototype so much effort. What I mean to say is that we should definitely decide on a best practice now and it sounds like most of the consensus is around JSON+DOJO, but we shouldn't be surprised if something much better comes up at some point in the future. BTW, as to how I see AJAX stuff fitting into OFBiz there are 2 big parts: 1. adding dynamic client side features to the form and screen (and other) widgets for things like dynamic lookups, expanding tree- browsers, movable screen-lets, etc. 2. special purpose pages that are meant to optimize a certain task, like the checkout stuff Tim mentioned, but probably even moreso for fancy back-end stuff For #1 replacing an AJAX framework is easy, but doesn't make as big a different in coding efficiency or anything, whereas for #2 something better would make coding easier, but require more rewriting... -David On Dec 13, 2006, at 4:53 PM, A. Zeneski wrote: > I am curious as to the pattern used to make the async calls, what > method > of return is used here? > > I see two possible patterns which would be useful: > > 1) JSON compressed objects. This will require more processing on the > front end to pull the data and format the HTML. I like leaving this in > the front end code personally. This is currently implemented using the > JSONServiceEventHandler request handler. > > 2) Screen Widget based HTML return. The Ajax request makes a simple > call > which returns a HTML fragment (developed using screen widgets). The > return is then just plugged in by updating the element content > (innerHTML). > > Either of these patterns is fine by me, and I believe that for most > applications a combination of the two would be necessary. > > I am sure there are many other ways to accomplish the same thing, but > since there is so much effort about to begin around 'Ajaxing' OFBiz > making sure things are consistent is #1 priority. > > I'm curious as to what other 'patterns' people have come up with > (if any). > > Andrew > > > Tim Ruppert wrote: >> Just to continue the push towards Dojo, I am about to submit a new >> Anonymous Checkout process for >> https://issues.apache.org/jira/browse/OFBIZ-510 which is based upon >> Dojo. You guys will like it - 3 steps and working towards my 2 step >> process goal. >> >> Cheers, >> Tim >> -- >> Tim Ruppert >> HotWax Media >> http://www.hotwaxmedia.com >> >> o:801.649.6594 >> f:801.649.6594 >> >> >> On Dec 13, 2006, at 4:36 PM, A. Zeneski wrote: >> >>> I used prototype for my initial test and found it very easy to work >>> with. Dojo on the other hand seemed to have a much broader learning >>> curve. It too me a while to get everything working, but I think >>> it is >>> due to the ajax part being more verbose. >>> >>> The extra features such as DND and other widgets is very nice as >>> well. >>> >>> The fact that DOJO is available under the BSD license is a plus >>> as well. >>> I think there should be no license issues with this library. >>> >>> Should we just say DOJO is the way to go for future Ajax based >>> tools in >>> OFBiz or do we need a formal vote? >>> >>> Andrew >>> >>> Alex D. Fleming wrote: >>>> Hi, >>>> >>>> Ooops Dojo is supported by IBM & AOL ....so I will prefer Dojo. >>>> >>>> http://dojotoolkit.org/foundation/ >>>> >>>> >>>> On 12/14/06, A. Zeneski <[hidden email]> wrote: >>>>> >>>>> Si (& all), >>>>> >>>>> I am in process of doing a lot of custom AJAX stuff for a new >>>>> client. I >>>>> intended on using prototype.js but based on conversations with >>>>> other >>>>> people decided to go with dojo instead. >>>>> >>>>> The main part of what I did was a JSON wrapper around the >>>>> service event >>>>> handler. What this does is allows you to define (controller) >>>>> requests >>>>> for ajax calls and have the results return in a lightweight (JSON) >>>>> format. >>>>> >>>>> This will work with any front end toolkit (Yahoo, Prototype, Dojo, >>>>> etc). >>>>> The main point is to define a standard pattern for AJAX >>>>> requests in >>>>> OFBiz. >>>>> >>>>> Deciding on a toolkit for the javascript is very important and >>>>> based on >>>>> what I have researched and discussed with other people, Dojo seems >>>>> to be >>>>> the way to go. >>>>> >>>>> Of course, I am open for discussion with this, but we really >>>>> should >>>>> decide on a standard for contributions back to OFBiz. The last >>>>> thing I >>>>> want to see is several different ways of communication and >>>>> different >>>>> toolkits being used. >>>>> >>>>> Prototype.js has been removed from SVN for the time being. As >>>>> soon as >>>>> there is code which uses one of the toolkits it can be added >>>>> back in. >>>>> >>>>> Let's start a community vote for the default AJAX toolkit to >>>>> use in >>>>> OFBiz. Based on the work I did, I have already setup the back end >>>>> pattern, services using JSON for the return. Now let's decide on a >>>>> front >>>>> end toolkit. >>>>> >>>>> First lets, define the candidates: >>>>> >>>>> 1) Dojo >>>>> 2) Prototype >>>>> >>>>> Andy >>>>> >> >> |
On Wednesday 13 December 2006 19:21, David E Jones wrote:
> BTW, as to how I see AJAX stuff fitting into OFBiz there are 2 big > parts: > > 1. adding dynamic client side features to the form and screen (and > other) widgets for things like dynamic lookups, expanding tree- > browsers, movable screen-lets, etc. > > 2. special purpose pages that are meant to optimize a certain task, > like the checkout stuff Tim mentioned, but probably even moreso for > fancy back-end stuff > > For #1 replacing an AJAX framework is easy, but doesn't make as big a > different in coding efficiency or anything, whereas for #2 something > better would make coding easier, but require more rewriting... We've had a few discussions going on over here about how to handle AJAX integration while still keeping security in mind. It seems like providing a limited dispatcher that only allows certain groups of services to be invoked is a good thing. Lots of services are already userLogin and Role protected, which is a good headstart. Maybe a specific AJAX dispatcher that has only services which are already vetted exposed through it? It would also be nice to have a high level facility for invoking those services and then triggering their results to something like the Dojo event framework when they return. With something like Comet these events could even be initiated by the server as if you had made a call on the client and then the delivery events occured. -- Ean Schuessler, CTO [hidden email] 214-720-0700 x 315 Brainfood, Inc. http://www.brainfood.com |
In reply to this post by David E Jones-2
Related to this, we have done some work already to create a dynamic
party lookup tag for the form widget. We were planning to wrap that up and put it on JIRA for everybody, but we may need to re-tweak it a bit in light of the jpublish/bsf changes. On Dec 13, 2006, at 5:21 PM, David E Jones wrote: > On an important but less pleasant note: the AJAX stuff we are > dealing with is very new in OFBiz, and outside of OFBiz the > "market" is still maturing quite a lot. I agree we should perhaps > vote on a best practice, though in general I'd rather see X task > implemented with DOJO required so much effort and with Prototype so > much effort. What I mean to say is that we should definitely decide > on a best practice now and it sounds like most of the consensus is > around JSON+DOJO, but we shouldn't be surprised if something much > better comes up at some point in the future. Best Regards, Si [hidden email] |
In reply to this post by Anil Patel-2
Nice, I like...
Anil Patel wrote: > 2) Screen Widget based HTML return. The Ajax request makes a simple call > which returns a HTML fragment (developed using screen widgets). The > return is then just plugged in by updating the element content (innerHTML). > > I have used this pattern with screen/ftl for generating html on demand. > > Regards > > > On 12/13/06, A. Zeneski <[hidden email]> wrote: >> >> I am curious as to the pattern used to make the async calls, what method >> of return is used here? >> >> I see two possible patterns which would be useful: >> >> 1) JSON compressed objects. This will require more processing on the >> front end to pull the data and format the HTML. I like leaving this in >> the front end code personally. This is currently implemented using the >> JSONServiceEventHandler request handler. >> >> 2) Screen Widget based HTML return. The Ajax request makes a simple call >> which returns a HTML fragment (developed using screen widgets). The >> return is then just plugged in by updating the element content >> (innerHTML). >> >> Either of these patterns is fine by me, and I believe that for most >> applications a combination of the two would be necessary. >> >> I am sure there are many other ways to accomplish the same thing, but >> since there is so much effort about to begin around 'Ajaxing' OFBiz >> making sure things are consistent is #1 priority. >> >> I'm curious as to what other 'patterns' people have come up with (if >> any). >> >> Andrew >> >> >> Tim Ruppert wrote: >> > Just to continue the push towards Dojo, I am about to submit a new >> > Anonymous Checkout process for >> > https://issues.apache.org/jira/browse/OFBIZ-510 which is based upon >> > Dojo. You guys will like it - 3 steps and working towards my 2 step >> > process goal. >> > >> > Cheers, >> > Tim >> > -- >> > Tim Ruppert >> > HotWax Media >> > http://www.hotwaxmedia.com >> > >> > o:801.649.6594 >> > f:801.649.6594 >> > >> > >> > On Dec 13, 2006, at 4:36 PM, A. Zeneski wrote: >> > >> >> I used prototype for my initial test and found it very easy to work >> >> with. Dojo on the other hand seemed to have a much broader learning >> >> curve. It too me a while to get everything working, but I think it is >> >> due to the ajax part being more verbose. >> >> >> >> The extra features such as DND and other widgets is very nice as well. >> >> >> >> The fact that DOJO is available under the BSD license is a plus as >> well. >> >> I think there should be no license issues with this library. >> >> >> >> Should we just say DOJO is the way to go for future Ajax based >> tools in >> >> OFBiz or do we need a formal vote? >> >> >> >> Andrew >> >> >> >> Alex D. Fleming wrote: >> >>> Hi, >> >>> >> >>> Ooops Dojo is supported by IBM & AOL ....so I will prefer Dojo. >> >>> >> >>> http://dojotoolkit.org/foundation/ >> >>> >> >>> >> >>> On 12/14/06, A. Zeneski <[hidden email]> wrote: >> >>>> >> >>>> Si (& all), >> >>>> >> >>>> I am in process of doing a lot of custom AJAX stuff for a new >> client. >> I >> >>>> intended on using prototype.js but based on conversations with other >> >>>> people decided to go with dojo instead. >> >>>> >> >>>> The main part of what I did was a JSON wrapper around the service >> event >> >>>> handler. What this does is allows you to define (controller) >> requests >> >>>> for ajax calls and have the results return in a lightweight (JSON) >> >>>> format. >> >>>> >> >>>> This will work with any front end toolkit (Yahoo, Prototype, Dojo, >> >>>> etc). >> >>>> The main point is to define a standard pattern for AJAX requests in >> >>>> OFBiz. >> >>>> >> >>>> Deciding on a toolkit for the javascript is very important and based >> on >> >>>> what I have researched and discussed with other people, Dojo seems >> >>>> to be >> >>>> the way to go. >> >>>> >> >>>> Of course, I am open for discussion with this, but we really should >> >>>> decide on a standard for contributions back to OFBiz. The last thing >> I >> >>>> want to see is several different ways of communication and different >> >>>> toolkits being used. >> >>>> >> >>>> Prototype.js has been removed from SVN for the time being. As >> soon as >> >>>> there is code which uses one of the toolkits it can be added back >> in. >> >>>> >> >>>> Let's start a community vote for the default AJAX toolkit to use in >> >>>> OFBiz. Based on the work I did, I have already setup the back end >> >>>> pattern, services using JSON for the return. Now let's decide on a >> >>>> front >> >>>> end toolkit. >> >>>> >> >>>> First lets, define the candidates: >> >>>> >> >>>> 1) Dojo >> >>>> 2) Prototype >> >>>> >> >>>> Andy >> >>>> >> > >> > >> >> >> > smime.p7s (4K) Download Attachment |
In reply to this post by David E Jones-2
Don't get me wrong, I'm not trying to limit us by any means. However,
there are currently a number of efforts going on in which Ajax will be used to enhance or update various pieces of the open source project. I for one would like to see an Ajax based packing screen go in the near future as well. What I mean by #1 priority, is to decide how we will standardize this [for now] during the first wave of development. If we have a handful of people all using different tools, we'll end up with a mess. To compare, we don't have some screens using Velocity and others using Freemarker, we have a standard for this. Even though Velocity is supported and people can use it in the custom solutions. Freemarker is standard for the open source contributions. Before we start getting patches from people, I think we need standards. This will save a lot of review time in the future. Andrew David E Jones wrote: > > Consistency is good, but I wouldn't say that's our #1 priority. We > definitely want primary and secondary best practices like we do with > other things, but also want to support a variety of technologies on the > framework level. In other words, if people want to do stuff in their own > code bases we don't want to block that, but yes for anything going into > the project it should follow a consistent set of patterns so that people > exploring the code for the first time will have a stream of "okay, okay, > okay, ..." going through their heads instead of a stream of "wtf, wtf, > wtf, ...". > > On an important but less pleasant note: the AJAX stuff we are dealing > with is very new in OFBiz, and outside of OFBiz the "market" is still > maturing quite a lot. I agree we should perhaps vote on a best practice, > though in general I'd rather see X task implemented with DOJO required > so much effort and with Prototype so much effort. What I mean to say is > that we should definitely decide on a best practice now and it sounds > like most of the consensus is around JSON+DOJO, but we shouldn't be > surprised if something much better comes up at some point in the future. > > BTW, as to how I see AJAX stuff fitting into OFBiz there are 2 big parts: > > 1. adding dynamic client side features to the form and screen (and > other) widgets for things like dynamic lookups, expanding tree-browsers, > movable screen-lets, etc. > > 2. special purpose pages that are meant to optimize a certain task, like > the checkout stuff Tim mentioned, but probably even moreso for fancy > back-end stuff > > For #1 replacing an AJAX framework is easy, but doesn't make as big a > different in coding efficiency or anything, whereas for #2 something > better would make coding easier, but require more rewriting... > > -David > > > On Dec 13, 2006, at 4:53 PM, A. Zeneski wrote: > >> I am curious as to the pattern used to make the async calls, what method >> of return is used here? >> >> I see two possible patterns which would be useful: >> >> 1) JSON compressed objects. This will require more processing on the >> front end to pull the data and format the HTML. I like leaving this in >> the front end code personally. This is currently implemented using the >> JSONServiceEventHandler request handler. >> >> 2) Screen Widget based HTML return. The Ajax request makes a simple call >> which returns a HTML fragment (developed using screen widgets). The >> return is then just plugged in by updating the element content >> (innerHTML). >> >> Either of these patterns is fine by me, and I believe that for most >> applications a combination of the two would be necessary. >> >> I am sure there are many other ways to accomplish the same thing, but >> since there is so much effort about to begin around 'Ajaxing' OFBiz >> making sure things are consistent is #1 priority. >> >> I'm curious as to what other 'patterns' people have come up with (if >> any). >> >> Andrew >> >> >> Tim Ruppert wrote: >>> Just to continue the push towards Dojo, I am about to submit a new >>> Anonymous Checkout process for >>> https://issues.apache.org/jira/browse/OFBIZ-510 which is based upon >>> Dojo. You guys will like it - 3 steps and working towards my 2 step >>> process goal. >>> >>> Cheers, >>> Tim >>> --Tim Ruppert >>> HotWax Media >>> http://www.hotwaxmedia.com >>> >>> o:801.649.6594 >>> f:801.649.6594 >>> >>> >>> On Dec 13, 2006, at 4:36 PM, A. Zeneski wrote: >>> >>>> I used prototype for my initial test and found it very easy to work >>>> with. Dojo on the other hand seemed to have a much broader learning >>>> curve. It too me a while to get everything working, but I think it is >>>> due to the ajax part being more verbose. >>>> >>>> The extra features such as DND and other widgets is very nice as well. >>>> >>>> The fact that DOJO is available under the BSD license is a plus as >>>> well. >>>> I think there should be no license issues with this library. >>>> >>>> Should we just say DOJO is the way to go for future Ajax based tools in >>>> OFBiz or do we need a formal vote? >>>> >>>> Andrew >>>> >>>> Alex D. Fleming wrote: >>>>> Hi, >>>>> >>>>> Ooops Dojo is supported by IBM & AOL ....so I will prefer Dojo. >>>>> >>>>> http://dojotoolkit.org/foundation/ >>>>> >>>>> >>>>> On 12/14/06, A. Zeneski <[hidden email]> wrote: >>>>>> >>>>>> Si (& all), >>>>>> >>>>>> I am in process of doing a lot of custom AJAX stuff for a new >>>>>> client. I >>>>>> intended on using prototype.js but based on conversations with other >>>>>> people decided to go with dojo instead. >>>>>> >>>>>> The main part of what I did was a JSON wrapper around the service >>>>>> event >>>>>> handler. What this does is allows you to define (controller) requests >>>>>> for ajax calls and have the results return in a lightweight (JSON) >>>>>> format. >>>>>> >>>>>> This will work with any front end toolkit (Yahoo, Prototype, Dojo, >>>>>> etc). >>>>>> The main point is to define a standard pattern for AJAX requests in >>>>>> OFBiz. >>>>>> >>>>>> Deciding on a toolkit for the javascript is very important and >>>>>> based on >>>>>> what I have researched and discussed with other people, Dojo seems >>>>>> to be >>>>>> the way to go. >>>>>> >>>>>> Of course, I am open for discussion with this, but we really should >>>>>> decide on a standard for contributions back to OFBiz. The last >>>>>> thing I >>>>>> want to see is several different ways of communication and different >>>>>> toolkits being used. >>>>>> >>>>>> Prototype.js has been removed from SVN for the time being. As soon as >>>>>> there is code which uses one of the toolkits it can be added back in. >>>>>> >>>>>> Let's start a community vote for the default AJAX toolkit to use in >>>>>> OFBiz. Based on the work I did, I have already setup the back end >>>>>> pattern, services using JSON for the return. Now let's decide on a >>>>>> front >>>>>> end toolkit. >>>>>> >>>>>> First lets, define the candidates: >>>>>> >>>>>> 1) Dojo >>>>>> 2) Prototype >>>>>> >>>>>> Andy >>>>>> >>> >>> > smime.p7s (4K) Download Attachment |
In reply to this post by Ean Schuessler
Ean,
When I implemented a test case using the new JSON service event handler I found that the requests behaved just like any other HTTP request. The security part is transparent, you require auth (but not HTTPS) for events which will be Ajax-Enabled. This uses the user login object found in the session (due to the fact that the url for the Ajax request is wrapped around <@ofbizUrl> tags) and the service engine handles the security. For screen based (non-JSON) responses, I think the security handling should be up to the page returned, just like any other page in OFBiz. Thoughts? Andrew Ean Schuessler wrote: > On Wednesday 13 December 2006 19:21, David E Jones wrote: >> BTW, as to how I see AJAX stuff fitting into OFBiz there are 2 big >> parts: >> >> 1. adding dynamic client side features to the form and screen (and >> other) widgets for things like dynamic lookups, expanding tree- >> browsers, movable screen-lets, etc. >> >> 2. special purpose pages that are meant to optimize a certain task, >> like the checkout stuff Tim mentioned, but probably even moreso for >> fancy back-end stuff >> >> For #1 replacing an AJAX framework is easy, but doesn't make as big a >> different in coding efficiency or anything, whereas for #2 something >> better would make coding easier, but require more rewriting... > > We've had a few discussions going on over here about how to handle AJAX > integration while still keeping security in mind. It seems like providing a > limited dispatcher that only allows certain groups of services to be invoked > is a good thing. Lots of services are already userLogin and Role protected, > which is a good headstart. Maybe a specific AJAX dispatcher that has only > services which are already vetted exposed through it? > > It would also be nice to have a high level facility for invoking those > services and then triggering their results to something like the Dojo event > framework when they return. With something like Comet these events could even > be initiated by the server as if you had made a call on the client and then > the delivery events occured. > smime.p7s (4K) Download Attachment |
I'd like to request before any votes go out that we see an actual implementation
of something useful in JSON and Dojo. I'm concerned that this approach is not lightweight enough for certain needs. First, it needs to support functionality that have a high rate of requests, such as autocomplete. This is critical. Second, the OFBiz AJAX system needs to be shallow and malleable. By shallow, I mean it should not have too many layers of complexity, abstraction or indirection. By malleable, I mean it should be easy to customize and build new effects or features for. Without these traits, it becomes difficult to develop and debug custom Ajax features. The reason I went with prototype was because of how easy it was to implement autocomplete in the form widget: It takes just two dozen lines of Javascript, some standard HTML form elements, and a standard Servlet method to process the autocomplete. And the requests are very fast, using the "ftl" handler. No special magic was required, just to load the required .js libraries. But this talk about having to write a JSON service event handler concerns me. If someone could demonstrate that it will be just as simple and straightforward, that would be great. - Leon A. Zeneski wrote: > Ean, > > When I implemented a test case using the new JSON service event handler > I found that the requests behaved just like any other HTTP request. The > security part is transparent, you require auth (but not HTTPS) for > events which will be Ajax-Enabled. This uses the user login object found > in the session (due to the fact that the url for the Ajax request is > wrapped around <@ofbizUrl> tags) and the service engine handles the > security. > > For screen based (non-JSON) responses, I think the security handling > should be up to the page returned, just like any other page in OFBiz. > > Thoughts? > > Andrew > > Ean Schuessler wrote: >> On Wednesday 13 December 2006 19:21, David E Jones wrote: >>> BTW, as to how I see AJAX stuff fitting into OFBiz there are 2 big >>> parts: >>> >>> 1. adding dynamic client side features to the form and screen (and >>> other) widgets for things like dynamic lookups, expanding tree- >>> browsers, movable screen-lets, etc. >>> >>> 2. special purpose pages that are meant to optimize a certain task, >>> like the checkout stuff Tim mentioned, but probably even moreso for >>> fancy back-end stuff >>> >>> For #1 replacing an AJAX framework is easy, but doesn't make as big a >>> different in coding efficiency or anything, whereas for #2 something >>> better would make coding easier, but require more rewriting... >> We've had a few discussions going on over here about how to handle AJAX >> integration while still keeping security in mind. It seems like providing a >> limited dispatcher that only allows certain groups of services to be invoked >> is a good thing. Lots of services are already userLogin and Role protected, >> which is a good headstart. Maybe a specific AJAX dispatcher that has only >> services which are already vetted exposed through it? >> >> It would also be nice to have a high level facility for invoking those >> services and then triggering their results to something like the Dojo event >> framework when they return. With something like Comet these events could even >> be initiated by the server as if you had made a call on the client and then >> the delivery events occured. >> |
Leon, check out https://issues.apache.org/jira/browse/OFBIZ-510 -
there's a VERY useful implementation up there and ready for review. This piece makes the anonymous checkout process usable - in my opinion. We don't actually need this piece for what we're doing at the moment since our customer only accepts credit cards for payment, but since I'm not in habit of just breaking other people's stuff and I don't really want to maintain tons of mods (other than straight custom work) outside of the OFBiz trunk, we took the time to do this. Please everyone take a look at this and let's get this initial AJAX direction decided upon sometime soon. We are building tons of apps based on a JSON / Dojo approach right now and would love for this to be adopted or know a good reason to change direction to something else. We're happy to do that if that's what we all decide. Cheers, Tim -- Tim Ruppert HotWax Media http://www.hotwaxmedia.com o:801.649.6594 f:801.649.6594 On Dec 13, 2006, at 8:39 PM, Leon Torres wrote: > I'd like to request before any votes go out that we see an actual > implementation of something useful in JSON and Dojo. I'm concerned > that this approach is not lightweight enough for certain needs. > > First, it needs to support functionality that have a high rate of > requests, such as autocomplete. This is critical. > > Second, the OFBiz AJAX system needs to be shallow and malleable. > By shallow, I mean it should not have too many layers of > complexity, abstraction or indirection. By malleable, I mean it > should be easy to customize and build new effects or features for. > Without these traits, it becomes difficult to develop and debug > custom Ajax features. > > The reason I went with prototype was because of how easy it was to > implement autocomplete in the form widget: It takes just two dozen > lines of Javascript, some standard HTML form elements, and a > standard Servlet method to process the autocomplete. And the > requests are very fast, using the "ftl" handler. No special magic > was required, just to load the required .js libraries. > > But this talk about having to write a JSON service event handler > concerns me. If someone could demonstrate that it will be just as > simple and straightforward, that would be great. > > - Leon > > > > A. Zeneski wrote: >> Ean, >> When I implemented a test case using the new JSON service event >> handler >> I found that the requests behaved just like any other HTTP >> request. The >> security part is transparent, you require auth (but not HTTPS) for >> events which will be Ajax-Enabled. This uses the user login object >> found >> in the session (due to the fact that the url for the Ajax request is >> wrapped around <@ofbizUrl> tags) and the service engine handles the >> security. >> For screen based (non-JSON) responses, I think the security handling >> should be up to the page returned, just like any other page in OFBiz. >> Thoughts? >> Andrew >> Ean Schuessler wrote: >>> On Wednesday 13 December 2006 19:21, David E Jones wrote: >>>> BTW, as to how I see AJAX stuff fitting into OFBiz there are 2 big >>>> parts: >>>> >>>> 1. adding dynamic client side features to the form and screen (and >>>> other) widgets for things like dynamic lookups, expanding tree- >>>> browsers, movable screen-lets, etc. >>>> >>>> 2. special purpose pages that are meant to optimize a certain task, >>>> like the checkout stuff Tim mentioned, but probably even moreso for >>>> fancy back-end stuff >>>> >>>> For #1 replacing an AJAX framework is easy, but doesn't make as >>>> big a >>>> different in coding efficiency or anything, whereas for #2 >>>> something >>>> better would make coding easier, but require more rewriting... >>> We've had a few discussions going on over here about how to >>> handle AJAX integration while still keeping security in mind. It >>> seems like providing a limited dispatcher that only allows >>> certain groups of services to be invoked is a good thing. Lots of >>> services are already userLogin and Role protected, which is a >>> good headstart. Maybe a specific AJAX dispatcher that has only >>> services which are already vetted exposed through it? >>> >>> It would also be nice to have a high level facility for invoking >>> those services and then triggering their results to something >>> like the Dojo event framework when they return. With something >>> like Comet these events could even be initiated by the server as >>> if you had made a call on the client and then the delivery events >>> occured. >>> |
In reply to this post by Leon Torres-2
Leon,
The JSON wrapper is for one purpose, to return service results in a lightweight compressed format for use by the Ajax javascript toolkit. This rather than XML is faster and more efficient as well as supported by all the toolkits I have see so far. This is only useful when you are looking to get specific values from a service call; and allows you to call existing services in your Ajax front end. One really nice way to use this would be for a simple upload widget. Call the service directly through a request event and have the results returned in an object ready for javascript to work with. This also uses the internal service security methods for authentication and validation. I think its a real clean way to integrate. I will try to put together a simple example and check it into one of the components, I will let you know when it is there. The other side is pulling in pre-formatted HTML. When doing this the HTML fragments should be configurable at least as much as the current templates are configurable. I like the idea of using screen definitions and FTL to generate this. Actually, this can be used for most things; but why when all you need to get back is an ID or some raw data. I think using a combination of these two methods (with any Ajax toolkit) is a great start. Andrew Leon Torres wrote: > I'd like to request before any votes go out that we see an actual > implementation of something useful in JSON and Dojo. I'm concerned that > this approach is not lightweight enough for certain needs. > > First, it needs to support functionality that have a high rate of > requests, such as autocomplete. This is critical. > > Second, the OFBiz AJAX system needs to be shallow and malleable. By > shallow, I mean it should not have too many layers of complexity, > abstraction or indirection. By malleable, I mean it should be easy to > customize and build new effects or features for. Without these traits, > it becomes difficult to develop and debug custom Ajax features. > > The reason I went with prototype was because of how easy it was to > implement autocomplete in the form widget: It takes just two dozen > lines of Javascript, some standard HTML form elements, and a standard > Servlet method to process the autocomplete. And the requests are very > fast, using the "ftl" handler. No special magic was required, just to > load the required .js libraries. > > But this talk about having to write a JSON service event handler > concerns me. If someone could demonstrate that it will be just as simple > and straightforward, that would be great. > > - Leon > > > > A. Zeneski wrote: >> Ean, >> >> When I implemented a test case using the new JSON service event handler >> I found that the requests behaved just like any other HTTP request. The >> security part is transparent, you require auth (but not HTTPS) for >> events which will be Ajax-Enabled. This uses the user login object found >> in the session (due to the fact that the url for the Ajax request is >> wrapped around <@ofbizUrl> tags) and the service engine handles the >> security. >> >> For screen based (non-JSON) responses, I think the security handling >> should be up to the page returned, just like any other page in OFBiz. >> >> Thoughts? >> >> Andrew >> >> Ean Schuessler wrote: >>> On Wednesday 13 December 2006 19:21, David E Jones wrote: >>>> BTW, as to how I see AJAX stuff fitting into OFBiz there are 2 big >>>> parts: >>>> >>>> 1. adding dynamic client side features to the form and screen (and >>>> other) widgets for things like dynamic lookups, expanding tree- >>>> browsers, movable screen-lets, etc. >>>> >>>> 2. special purpose pages that are meant to optimize a certain task, >>>> like the checkout stuff Tim mentioned, but probably even moreso for >>>> fancy back-end stuff >>>> >>>> For #1 replacing an AJAX framework is easy, but doesn't make as big a >>>> different in coding efficiency or anything, whereas for #2 something >>>> better would make coding easier, but require more rewriting... >>> We've had a few discussions going on over here about how to handle >>> AJAX integration while still keeping security in mind. It seems like >>> providing a limited dispatcher that only allows certain groups of >>> services to be invoked is a good thing. Lots of services are already >>> userLogin and Role protected, which is a good headstart. Maybe a >>> specific AJAX dispatcher that has only services which are already >>> vetted exposed through it? >>> >>> It would also be nice to have a high level facility for invoking >>> those services and then triggering their results to something like >>> the Dojo event framework when they return. With something like Comet >>> these events could even be initiated by the server as if you had made >>> a call on the client and then the delivery events occured. >>> smime.p7s (4K) Download Attachment |
In reply to this post by Si Chen-2
Dear all,
Nice read... can help Ofbiz decide or atleast debate what they need. Bad thing is there is no section for XAL. Exploring Ajax Runtime Offerings http://dev2dev.bea.com/pub/a/2006/11/exploring-ajax.html http://dev2dev.bea.com/lpt/a/542 Nexaweb had jRex and help XAL - Apache Project. ( jRex is great , but could not see on the site anymore :( ). My pick for Ajax (1) Declarative (2) Java Based (3) Server Side (4) Open Source :) Nice to support Compound Document ... Helps evolutionary adding new tags and HTML code is reused. Chand ----- Original Message ----- From: "Alex D. Fleming" <[hidden email]> To: <[hidden email]> Sent: Wednesday, December 13, 2006 1:07 PM Subject: Re: json + prototype + ajax? > Hi, > > Ooops Dojo is supported by IBM & AOL ....so I will prefer Dojo. > > http://dojotoolkit.org/foundation/ > > >> >> What we are doing is >> >> (1) Write code in xal/ nxml format. >> (2) reUse the screen freemaker code >> (3) Merge both of them. >> >> e.g. registration ; login ; >> >> help people to convert HTML to XAL + HTML in few screens. >> >> >> Chand >> >> >> >> >>> Sorry, >>> >>> For not getting back to you right away. I had my spam >> filter set to >>> high and it blocked this message. >>> >>> Currently I don't believe any one has used either XAP or >> Nexaweb with >>> OFBiz. That being said, XAP and Nexaweb could be used >> easily. I could >>> see a couple places for it to be used: 1.) On the >> management of the >>> eCommerce 2.) In the actual eCommerce front end. >>> >>> I would like to help with working out who we (XAP and >> OFBiz) could >>> collaborate in some projects. >>> >>> If users want to enhance features of OFBiz with XAP >> functionality, a >>> good way to start would be to use the xModify portion of >> XAP, to modify >>> the HTML DOM asynchronously. This is very easy to use. >> I have written >>> up an explanation of how to do this. >>> >>> http://www.rockstarapps.com/pmwiki/pmwiki.php? >> n=XAP.Rssreader >>> >>> XAP widget library may also be helpful and there are >> sample on the xap >>> website, on how to use this. >>> >>> http://incubator.apache.org/xap >>> >>> >>> >>> Bob (Buffone) >>> >>> -----Original Message----- >>> From: Chandresh Turakhia >> [mailto:[hidden email]] >>> Sent: Tuesday, November 28, 2006 4:32 AM >>> To: [hidden email] >>> Cc: [hidden email] >>> Subject: XAP ( Nexaweb ) + Ofbiz >>> >>> Hi , >>> >>> Greetings... >>> >>> Did anyone work with Ofbiz + XAP or OFbiz + Nexaweb. >>> >>> Any pointers welcome . >>> >>> Thanks in advance. >>> >>> Chand >>> >> >> >> ---------------------------------------------------------- >> http://www.bhartitelesoft.com >> Telecommunication and E-Commerce solutions for the future. >> >> >> >> >> > |
Administrator
|
In reply to this post by cjhowe
I don't know much about Ajax and Dojo and such. But IMHO this is a very important point.
Chris (or anybody aware ;o) do you know how exactly this is achieved ? Thanks Jacques From: "Chris Howe" <[hidden email]> > I'm not sure about the other toolkits, but when I > looked around for some implementation ideas, I like > the fact that one of Dojo's precepts is to degrade > when not JS is turned off. This may be true of other > toolkits as well, but Dojo states it specifically. |
I may think that OFBiz widgets compatibility is a very important point.
That's I think that declarative ajax framework would be of simplier integration "Jacques Le Roux" a écrit le 14/12/2006 10:18 : > I don't know much about Ajax and Dojo and such. But IMHO this is a very important point. > Chris (or anybody aware ;o) do you know how exactly this is achieved ? > > Thanks > > Jacques > > From: "Chris Howe" <[hidden email]> > >> I'm not sure about the other toolkits, but when I >> looked around for some implementation ideas, I like >> the fact that one of Dojo's precepts is to degrade >> when not JS is turned off. This may be true of other >> toolkits as well, but Dojo states it specifically. >> > > |
In reply to this post by Jacques Le Roux
Also in regards to the comments in OFBIZ-510 :
The ability to degrade doesn't solve itself simply because you're using Dojo or any other toolkit that is designed to degrade. You still need to have the functionality on your page. Degrading essentially means that the javascript simply increases functionality, the javascript isn't replacing functionality. Dojo does this because the javascript actions are identified either by css classes or by additional attributes in the tag that mean nothing to the browser that hasn't received any instruction from javascript. There are a lot of places that utilize javascript in OFBiz. For instance the calendar popups. The JS degrades here because you can still enter date/time information without the popup. On the JIRA page, the functionality of adding a comment doesn't degrade and I'm sure the people at Atlassian would like to know that. This appears to be based on their custom JS additions, not on the parts of OFBiz they are using. In a checkout process you would degrade the JS by having form submits that actually go to the next step, but have the JS disable that submit button. By doing that if there is JS, the form doesn't get submitted in the traditional manner but instead the JS handles the next sequence. If JS is disabled, then there is nothing there to disable the submit button and the form information gets sent for processing like it always has. This is a great place for an AJAX type solution because the checkout logic is based only on updating the ShoppingCart object and not on entering information in the database. --- Jacques Le Roux <[hidden email]> wrote: > I don't know much about Ajax and Dojo and such. But > IMHO this is a very important point. > Chris (or anybody aware ;o) do you know how exactly > this is achieved ? > > Thanks > > Jacques > > From: "Chris Howe" <[hidden email]> > > I'm not sure about the other toolkits, but when I > > looked around for some implementation ideas, I > like > > the fact that one of Dojo's precepts is to degrade > > when not JS is turned off. This may be true of > other > > toolkits as well, but Dojo states it specifically. > |
Administrator
|
Thanks Chris,
This is really clear to me now :o) Jacques ----- Original Message ----- From: "Chris Howe" <[hidden email]> To: <[hidden email]>; "Jacques Le Roux" <[hidden email]> Sent: Thursday, December 14, 2006 3:56 PM Subject: Re: json + prototype + ajax? > Also in regards to the comments in OFBIZ-510 : > > The ability to degrade doesn't solve itself simply > because you're using Dojo or any other toolkit that is > designed to degrade. You still need to have the > functionality on your page. > > Degrading essentially means that the javascript simply > increases functionality, the javascript isn't > replacing functionality. Dojo does this because the > javascript actions are identified either by css > classes or by additional attributes in the tag that > mean nothing to the browser that hasn't received any > instruction from javascript. > > There are a lot of places that utilize javascript in > OFBiz. For instance the calendar popups. The JS > degrades here because you can still enter date/time > information without the popup. On the JIRA page, the > functionality of adding a comment doesn't degrade and > I'm sure the people at Atlassian would like to know > that. This appears to be based on their custom JS > additions, not on the parts of OFBiz they are using. > > In a checkout process you would degrade the JS by > having form submits that actually go to the next step, > but have the JS disable that submit button. By doing > that if there is JS, the form doesn't get submitted in > the traditional manner but instead the JS handles the > next sequence. If JS is disabled, then there is > nothing there to disable the submit button and the > form information gets sent for processing like it > always has. > > This is a great place for an AJAX type solution > because the checkout logic is based only on updating > the ShoppingCart object and not on entering > information in the database. > > > > --- Jacques Le Roux <[hidden email]> > wrote: > > > I don't know much about Ajax and Dojo and such. But > > IMHO this is a very important point. > > Chris (or anybody aware ;o) do you know how exactly > > this is achieved ? > > > > Thanks > > > > Jacques > > > > From: "Chris Howe" <[hidden email]> > > > I'm not sure about the other toolkits, but when I > > > looked around for some implementation ideas, I > > like > > > the fact that one of Dojo's precepts is to degrade > > > when not JS is turned off. This may be true of > > other > > > toolkits as well, but Dojo states it specifically. > > |
Administrator
|
In reply to this post by Chandresh Turakhia
Hi all,
I found this article very interesting notably the Client-centric vs. server-centric approach and in this the table comparing Ajax frameworks. I tried Echo2 demo, a new world ? Jacques ----- Original Message ----- From: "Chandresh Turakhia" <[hidden email]> To: <[hidden email]> Cc: <[hidden email]>; "Bob Buffone" <[hidden email]>; "Chandresh Turakhia" <[hidden email]>; <[hidden email]> Sent: Thursday, December 14, 2006 9:49 AM Subject: Re: json + prototype + ajax? Dear all, Nice read... can help Ofbiz decide or atleast debate what they need. Bad thing is there is no section for XAL. Exploring Ajax Runtime Offerings http://dev2dev.bea.com/pub/a/2006/11/exploring-ajax.html http://dev2dev.bea.com/lpt/a/542 Nexaweb had jRex and help XAL - Apache Project. ( jRex is great , but could not see on the site anymore :( ). My pick for Ajax (1) Declarative (2) Java Based (3) Server Side (4) Open Source :) Nice to support Compound Document ... Helps evolutionary adding new tags and HTML code is reused. Chand ----- Original Message ----- From: "Alex D. Fleming" <[hidden email]> To: <[hidden email]> Sent: Wednesday, December 13, 2006 1:07 PM Subject: Re: json + prototype + ajax? > Hi, > > Ooops Dojo is supported by IBM & AOL ....so I will prefer Dojo. > > http://dojotoolkit.org/foundation/ > > >> >> What we are doing is >> >> (1) Write code in xal/ nxml format. >> (2) reUse the screen freemaker code >> (3) Merge both of them. >> >> e.g. registration ; login ; >> >> help people to convert HTML to XAL + HTML in few screens. >> >> >> Chand >> >> >> >> >>> Sorry, >>> >>> For not getting back to you right away. I had my spam >> filter set to >>> high and it blocked this message. >>> >>> Currently I don't believe any one has used either XAP or >> Nexaweb with >>> OFBiz. That being said, XAP and Nexaweb could be used >> easily. I could >>> see a couple places for it to be used: 1.) On the >> management of the >>> eCommerce 2.) In the actual eCommerce front end. >>> >>> I would like to help with working out who we (XAP and >> OFBiz) could >>> collaborate in some projects. >>> >>> If users want to enhance features of OFBiz with XAP >> functionality, a >>> good way to start would be to use the xModify portion of >> XAP, to modify >>> the HTML DOM asynchronously. This is very easy to use. >> I have written >>> up an explanation of how to do this. >>> >>> http://www.rockstarapps.com/pmwiki/pmwiki.php? >> n=XAP.Rssreader >>> >>> XAP widget library may also be helpful and there are >> sample on the xap >>> website, on how to use this. >>> >>> http://incubator.apache.org/xap >>> >>> >>> >>> Bob (Buffone) >>> >>> -----Original Message----- >>> From: Chandresh Turakhia >> [mailto:[hidden email]] >>> Sent: Tuesday, November 28, 2006 4:32 AM >>> To: [hidden email] >>> Cc: [hidden email] >>> Subject: XAP ( Nexaweb ) + Ofbiz >>> >>> Hi , >>> >>> Greetings... >>> >>> Did anyone work with Ofbiz + XAP or OFbiz + Nexaweb. >>> >>> Any pointers welcome . >>> >>> Thanks in advance. >>> >>> Chand >>> >> >> >> ---------------------------------------------------------- >> http://www.bhartitelesoft.com >> Telecommunication and E-Commerce solutions for the future. >> >> >> >> >> > |
In reply to this post by Andrew Zeneski
Here are my two cents about this interesting thread:
1) the Ajax toolkit's license must be fully compatible with the OFBiz license 2) even if it's a good thing to try to find one official Ajax toolkit for OFBiz now and finally get this ball running, I think we should do this but also review the decision (and the results we'll get with the adopted framework) in 2-4 months from now and possibly return on it; I mean that we should keep an open-minded approach and also consider new solutions (or criticism to the adopted toolkit) since I think that in the Ajax world the effects of the 'software darwinism' (http://www.apache.org/foundation/glossary.html#SoftwareDarwinism) still are not mature Jacopo A. Zeneski wrote: > I'll let this thread run a little while longer before we say to have a > full official vote. As of right now, it appears most people are looking > at Dojo, and that is fine with me. > > It seems that most of these toolkits do the same thing, so to me its > just a matter of making a decision so I can push forward with my work. > > As for being more active on the lists, sorry I have been MIA for so > long. I've been involved in a lot of custom (non open source) work as of > late and apologize for not being around. I will do my best to be here as > much as possible. Thanks! > > Andrew > |
Administrator
|
I totally agree with these POV !
Jacques ----- Original Message ----- From: "Jacopo Cappellato" <[hidden email]> To: <[hidden email]> Sent: Friday, December 15, 2006 8:07 AM Subject: Re: json + prototype + ajax? > Here are my two cents about this interesting thread: > > 1) the Ajax toolkit's license must be fully compatible with the OFBiz > license > 2) even if it's a good thing to try to find one official Ajax toolkit > for OFBiz now and finally get this ball running, I think we should do > this but also review the decision (and the results we'll get with the > adopted framework) in 2-4 months from now and possibly return on it; I > mean that we should keep an open-minded approach and also consider new > solutions (or criticism to the adopted toolkit) since I think that in > the Ajax world the effects of the 'software darwinism' > (http://www.apache.org/foundation/glossary.html#SoftwareDarwinism) still > are not mature > > Jacopo > > A. Zeneski wrote: > > I'll let this thread run a little while longer before we say to have a > > full official vote. As of right now, it appears most people are looking > > at Dojo, and that is fine with me. > > > > It seems that most of these toolkits do the same thing, so to me its > > just a matter of making a decision so I can push forward with my work. > > > > As for being more active on the lists, sorry I have been MIA for so > > long. I've been involved in a lot of custom (non open source) work as of > > late and apologize for not being around. I will do my best to be here as > > much as possible. Thanks! > > > > Andrew > > |
In reply to this post by Jacopo Cappellato
I would certainly love to see the highly useful Anonymous Checkout
Process example used in OFBiz - at least until someone else comes up with a good reason to remove Dojo and go to a different front end framewok. Does anyone have any real objections to doing this in light of the fact that no one has another example _and_ that the checkout process is unnecessarily tedious? Cheers, Tim -- Tim Ruppert HotWax Media http://www.hotwaxmedia.com o:801.649.6594 f:801.649.6594 On Dec 15, 2006, at 12:07 AM, Jacopo Cappellato wrote: > Here are my two cents about this interesting thread: > > 1) the Ajax toolkit's license must be fully compatible with the > OFBiz license > 2) even if it's a good thing to try to find one official Ajax > toolkit for OFBiz now and finally get this ball running, I think we > should do this but also review the decision (and the results we'll > get with the adopted framework) in 2-4 months from now and possibly > return on it; I mean that we should keep an open-minded approach > and also consider new solutions (or criticism to the adopted > toolkit) since I think that in the Ajax world the effects of the > 'software darwinism' (http://www.apache.org/foundation/ > glossary.html#SoftwareDarwinism) still are not mature > > Jacopo > > A. Zeneski wrote: >> I'll let this thread run a little while longer before we say to >> have a >> full official vote. As of right now, it appears most people are >> looking >> at Dojo, and that is fine with me. >> It seems that most of these toolkits do the same thing, so to me its >> just a matter of making a decision so I can push forward with my >> work. >> As for being more active on the lists, sorry I have been MIA for so >> long. I've been involved in a lot of custom (non open source) work >> as of >> late and apologize for not being around. I will do my best to be >> here as >> much as possible. Thanks! >> Andrew |
Tim,
To your post http://issues.apache.org/jira/browse/OFBIZ-510#action_12458496 I would want to see it degrade. The fact that SVN OFBiz does not degrade should not be support to introduce more code that follows a poor pattern. The current state of OFBiz not being able to add to cart when JS is turned off needs to be reported as a bug and fixed. Wanting to see this degrade should especially be true for a functionality of "Anonymous" checkout. This means you're wanting to sell product to random people. So, if you're wanting to sell to random people, you would want to lower the barriers as far as possible for them to use your site. Good web surfing practice is to white list JS for sites that you trust. This is especially true in corporate environments. So to summarize, my two cents would be to report the add to cart bug to JIRA, fix OFBiz's SVN of this pattern and degrade the JS in the anonymous checkout patch. But to quickly see this kind of functionality into OFBiz, I would prefer to see it added to a sandbox so that others could help work out these peculiarities. --- Tim Ruppert <[hidden email]> wrote: > I would certainly love to see the highly useful > Anonymous Checkout > Process example used in OFBiz - at least until > someone else comes up > with a good reason to remove Dojo and go to a > different front end > framewok. Does anyone have any real objections to > doing this in > light of the fact that no one has another example > _and_ that the > checkout process is unnecessarily tedious? > > Cheers, > Tim > -- > Tim Ruppert > HotWax Media > http://www.hotwaxmedia.com > > o:801.649.6594 > f:801.649.6594 > > > On Dec 15, 2006, at 12:07 AM, Jacopo Cappellato > wrote: > > > Here are my two cents about this interesting > thread: > > > > 1) the Ajax toolkit's license must be fully > compatible with the > > OFBiz license > > 2) even if it's a good thing to try to find one > official Ajax > > toolkit for OFBiz now and finally get this ball > running, I think we > > should do this but also review the decision (and > the results we'll > > get with the adopted framework) in 2-4 months from > now and possibly > > return on it; I mean that we should keep an > open-minded approach > > and also consider new solutions (or criticism to > the adopted > > toolkit) since I think that in the Ajax world the > effects of the > > 'software darwinism' > (http://www.apache.org/foundation/ > > glossary.html#SoftwareDarwinism) still are not > mature > > > > Jacopo > > > > A. Zeneski wrote: > >> I'll let this thread run a little while longer > before we say to > >> have a > >> full official vote. As of right now, it appears > most people are > >> looking > >> at Dojo, and that is fine with me. > >> It seems that most of these toolkits do the same > thing, so to me its > >> just a matter of making a decision so I can push > forward with my > >> work. > >> As for being more active on the lists, sorry I > have been MIA for so > >> long. I've been involved in a lot of custom (non > open source) work > >> as of > >> late and apologize for not being around. I will > do my best to be > >> here as > >> much as possible. Thanks! > >> Andrew > > |
Chris, JavaScript is ingrained in just about anything that submits
forms anymore. I could understand saying something like this about Flash - however my mind is changing on that as well - but saying that you have to be able to manage something as complicated as the feature set that OFBiz employs without JavaScript is almost like saying - not everyone is off of Netscape 4.x and we need to set the bar _that_ low for our CSS/HTML standardization. My vote for this is to move forward, not hang back. JavaScript, as much as I'm not a huge fan, is a reality in todays world - and the use of of DOES NOT make the code bug filled. Cheers, Tim -- Tim Ruppert HotWax Media http://www.hotwaxmedia.com o:801.649.6594 f:801.649.6594 On Dec 15, 2006, at 8:07 AM, Chris Howe wrote: > Tim, > > To your post > http://issues.apache.org/jira/browse/OFBIZ-510#action_12458496 > > I would want to see it degrade. The fact that SVN > OFBiz does not degrade should not be support to > introduce more code that follows a poor pattern. The > current state of OFBiz not being able to add to cart > when JS is turned off needs to be reported as a bug > and fixed. > > Wanting to see this degrade should especially be true > for a functionality of "Anonymous" checkout. This > means you're wanting to sell product to random people. > So, if you're wanting to sell to random people, you > would want to lower the barriers as far as possible > for them to use your site. Good web surfing practice > is to white list JS for sites that you trust. This is > especially true in corporate environments. > > So to summarize, my two cents would be to report the > add to cart bug to JIRA, fix OFBiz's SVN of this > pattern and degrade the JS in the anonymous checkout > patch. But to quickly see this kind of functionality > into OFBiz, I would prefer to see it added to a > sandbox so that others could help work out these > peculiarities. > > > --- Tim Ruppert <[hidden email]> wrote: > >> I would certainly love to see the highly useful >> Anonymous Checkout >> Process example used in OFBiz - at least until >> someone else comes up >> with a good reason to remove Dojo and go to a >> different front end >> framewok. Does anyone have any real objections to >> doing this in >> light of the fact that no one has another example >> _and_ that the >> checkout process is unnecessarily tedious? >> >> Cheers, >> Tim >> -- >> Tim Ruppert >> HotWax Media >> http://www.hotwaxmedia.com >> >> o:801.649.6594 >> f:801.649.6594 >> >> >> On Dec 15, 2006, at 12:07 AM, Jacopo Cappellato >> wrote: >> >>> Here are my two cents about this interesting >> thread: >>> >>> 1) the Ajax toolkit's license must be fully >> compatible with the >>> OFBiz license >>> 2) even if it's a good thing to try to find one >> official Ajax >>> toolkit for OFBiz now and finally get this ball >> running, I think we >>> should do this but also review the decision (and >> the results we'll >>> get with the adopted framework) in 2-4 months from >> now and possibly >>> return on it; I mean that we should keep an >> open-minded approach >>> and also consider new solutions (or criticism to >> the adopted >>> toolkit) since I think that in the Ajax world the >> effects of the >>> 'software darwinism' >> (http://www.apache.org/foundation/ >>> glossary.html#SoftwareDarwinism) still are not >> mature >>> >>> Jacopo >>> >>> A. Zeneski wrote: >>>> I'll let this thread run a little while longer >> before we say to >>>> have a >>>> full official vote. As of right now, it appears >> most people are >>>> looking >>>> at Dojo, and that is fine with me. >>>> It seems that most of these toolkits do the same >> thing, so to me its >>>> just a matter of making a decision so I can push >> forward with my >>>> work. >>>> As for being more active on the lists, sorry I >> have been MIA for so >>>> long. I've been involved in a lot of custom (non >> open source) work >>>> as of >>>> late and apologize for not being around. I will >> do my best to be >>>> here as >>>> much as possible. Thanks! >>>> Andrew >> >> |
Free forum by Nabble | Edit this page |