Thursday, September 27th, 2007

Be wary of Ajax! (So says Oracle)

Category: Ajax, Java

<p>I found it kind of ironic that Ted Farrell, Oracle’s chief architect and vice president of tools and middleware, warned many AJAXWorld conference attendees to be wary of Ajax. So let me get this straight. You go to an Ajax-centric conference to tell the developers who are there to learn about Ajax-related technologies to be wary of the very topic they’re there to learn about. Hmmmmmmm.

Ted Farrell, the exec setting technical and strategic direction of Oracle’s own development tools, told The Register that most AJAX frameworks focus on the interface and pay lip service to the most challenging part – integration with back-end servers.

Adopting AJAX-based rich interfaces and internet-based applications at this stage could leave their organizations stranded in the future, lacking either skills or an upgrade path to continue working, he argues.

He does have a point in that most of the current frameworks are really focused on the client-side of things and yes, it would certainly be nice to have some tighter integration with backend systems. But with developers taking abstraction seriously and APIs being developed to simplify access to backend systems via JavaScript, I’m not 100% convinced of Mr. Farrell’s conclusions that organizations will get left out in the cold.

Now, considering Oracle’s commitment to Java, I don’t think you’ll be surprised by the fact that they’re pushing Java Server Faces as the solution for developing rich UIs:

According to Farrell, JSF already delivers the ability to build a rich Web 2.0-style interface while abstracting the developer from integration with the back end. “JSF separates the object model from rendering, you get events back and a render kit decides how the event is rendered,” he said.

He also takes a swipe at other RIA technologies including Adobe AIR & Silverlight (and I’m assuming Adobe Flex as well) basically saying that they provide no standard method of integrating with backend systems.

So is JSF the silver bullet we’ve all been waiting for? Should we give up on Dojo, Prototype, Silverlight or Flex?

I’m not jumping ship but I’d definitely like to hear what you think.

Related Content:

Posted by Rey Bango at 6:30 am
47 Comments

+++--
3.4 rating from 42 votes

47 Comments »

Comments feed TrackBack URI

Well … my personal oppinion is that there is room for everybody, even for jsf. My guess is that mr. Farrell is frustrated by the small scale adoption of jsf (released 2004) versus struts or whatever template/view technologies in the java world.
But hey, the internet is not spinning arround java or java based technologies.
So what if the ajax frameworks provide no standard method for integrating with the backend (actualy it’s all http, so it’s a kind of standard …), who says that everybody should do what jsf does ? Since when jsf is THE standard for the web ?
By not providing a “standard method” to interact with the backend, the developers can define the interaction depending on the purpose (why using SOAP for a simple rating system, when you can send a plain POST and receive JSON ?).
The beauty of the web programming is it’s diversity of technologies, each suited for a specific purpose. It’s not about ruling the world, it’s about living inside it.

Comment by jsf — September 27, 2007

I think that mr-oracle is just scared that people will stop using their tighty integrated over-abstracted stuff and figure out easier way to accomplish the same thing without 5-figure oracle licenses…

In dutch we have a special saying: “Wij van wc-eend adviseren wc-eend” wich translates to something like: “We from Oracle advice everybody to use Oracle!”

Comment by SchizoDuckie — September 27, 2007

That’s pretty funny, so his argument is AJAX is a set client side technologies!?

The whole point and a big advantage to businesses who won’t get “stranded” locked down into 1 specific environment. It separates the client side logic vs. server side logic. You can switch your whole backend without touching the client side or change the client side without touching the backend.

So this is bad news for oracle, sap and most of the corporate IT sector
who try to lock their customers into 1 particular technology. Thank goodness there’s google who actively promotes open standards and technologies.

Comment by Jonathan — September 27, 2007

As a Java- (and esp. J2EE-) hater and Python-enthusiast I highly doubt JSF will become an option for me. Same goes for hundreds of billions of Ruby-, Perl- and PHP-programmers, I guess.

Java used to be a great, promising platform in the late 90s, but these days there’s so many great frameworks to build your webapps upon, that help you reaching your targets much faster and easier.

Comment by Daniel Haus — September 27, 2007

I went to a talk by Ted Farell recently (http://www.it-eye.nl/weblog/2007/09/20/oracle-java-tools-technical-briefing/) and his point was that writing ajax apps yourself is too difficult even with most ajax libraries. By handcoding all the javascript/ajax parts you’re creating a maintenance nightmare. Developers should look for ui components which handle the ajax details for the developers. The components should handle all the browser specific details.

I think oracle’s is doing a nice job with adf faces/jdeveloper. Using jdeveloper you can more or less generated/visually develop ajax applications without writing javascript. Pick a database table, generate a data control, drop the data control on a jsf page, and you have an ajax enhanced table. The adf faces (jsf) components in jdeveloper 11 are pretty slick, declarative drag and drop, requerying data on the fly while pulling down a scrollbar, all without writing javascript code.

Regarding flex he said that it was a closed box, that didn’t allow you to integrate different applications. In jdeveloper 11g you’ll be able to add panels to any jsf page after deployment. Sort of like adding portlets to a portal page, but in this case for any jsf page you have. This is implemented using webcenter and oracle metadata services (MDS), which allows you to transform any deployed declarative piece of code (jsf, xml, xslt, bpel, esb, etc). This will enable you to customize pages for users or departments, without having to be build in the original application. The big advantage of this approach is that you can modify boxed applications without having to modify the original code. Less upgrade nightmares when you get a new release of the purchased application.

Comment by Andrej — September 27, 2007

“He does have a point in that most of the current frameworks are really focused on the client-side of things and yes, it would certainly be nice to have some tighter integration with backend systems.”

Uh, no. IMHO, Ajax applications should be loosely coupled to back-end services. And they certainly should *not* be bound to any single back-end server language. Ajax apps are most flexible when they are simply consumers of web services.

Comment by Mark P — September 27, 2007

It’s funny that he tries to shoot flex down due to it’s lack of support with server side even thrugh flex can communicate with ANY serverside service, you just to tell it how too.

Is that not what we do with ANY language?

Comment by Jack Gleeson — September 27, 2007

I don’t know why he’s so scared of Javascript – maybe because Oracle can’t sell it. Fact is. once you get the hang of JS a lot of things become really routine and predictable. And it does focus attention away from the back end, which is bad news for Oracle…

If they really wanted to be useful in this debate, they might push harder for open standards.

Comment by david — September 27, 2007

ROFLOL && ROFLMAO…!!!! :))))
After being finished with all of my laughing I’d like to pinpoint a couple of interesting observations. First Ted Farrell obviously has an agenda with what he says. He’s trying to sell (one of) the “alternative” solution(s).
Second I must say I really admire not only his view but also his courage…
Quite a courageous thing to do in fact (warning against “JavaScript inclusion libraries” at the AjaxWorld Conference where probably 98% of the participants have either created their own “JavaScript inclusion library” or at least consumed several of those beasts themselves)
But most importantly, he’s BULLS EYE…!!
To perceive JavaScript as anything more than the “glue” that binds the user experience together with the server and therefore the job of a “compiler” (read; Framework) is not only incredible stupid but also probably directly dangerous for both you and your employer…
JavaScript is the “New Assembly Language” and the job of the “New Compiler” (read; Framework) to take care of. Now you don’t want to write your next Enterprise Solution in Assembly code and you certainly DON’T want to create it in JavaScript. There’s plenty of good Ajax Libraries out there (http://ajaxwidgets.com being _one_ of them) that completely abstracts away JavaScript to the places where it belongs; In the heads of the framework developers and NOT in the heads of application developers.

Comment by Thomas Hansen — September 27, 2007

JSF couldn’t deliver (for reasons I don’t quite understand) and we should let it die in piece. On the other hand, he does have a point about the chaos in the AJAX tools/libraries/frameworks, but this is something to be expected with emerging technologies.

I personally like the idea of having my data in a database, a middle layer that exposes it as XML (written in Java, .NET or whatever…) and a RIA client as GUI. This architecture fits mostly consumer and not business apps and this is why Oracle doesn’t get AJAX.

Comment by Jacob — September 27, 2007

JSF is a serverside technology. This puts the RIA load on the server, making it inherently less scalable than client side technologies and creating higher bandwidth requirements during operation.

Because the database can often be the bottleneck for data driven applications, corporate IT will usually ensure that high-end servers are available to support the RDBMS and these may well be shared with the web application. But in other environments, ideally it is better to use the abundant client side CPU cycles that are scalable with the number of clients and reserve the server horsepower for tasks that must be located at the service source.

Comment by Alex Tolley — September 27, 2007

Sun: The Network is the Computer
Oracle: The database is the application
Microsoft: Your potential, Our forced upgrades
Ajaxians: My library can beat-up your library’s dad

Comment by James MacFarlane — September 27, 2007

@James: Awesome! LOL!

Comment by Rey Bango — September 27, 2007

“There’s plenty of good Ajax Libraries out there that completely abstracts away JavaScript to the places where it belongs;”

The problem is most of these “frameworks” are not really frameworks. And when they are merely toolkits, sometimes they overstep their bounds and try to act as frameworks (poorly).

The state of the toolkits/frameworks is not nearly mature enough to claim that the javascript can be abstracted to ‘places where it belongs’.

As far as “Ajax” goes, it just means messaging. An “ajax” framework should be about 100 lines of code. A DHTML GUI toolkit is a different beast entirely.

Nate Grover

Comment by Nate Grover — September 27, 2007

Tighter integration with the backend? I suppose that would be nice, but I really only need the datasource so I could care less. I prefer the loose coupling anyways because then the client-side code is more reusable.

Comment by Andy Kant — September 27, 2007

I saw one of the presentations of JSF at AJAXWorld, and Java is my primary server language, but the JSF talk felt like someone proclaiming the power of the titanic as it is sinking.

Comment by Kris Zyp — September 27, 2007

I went to about 5 minutes of this talk, and after listening to a number of factually incorrect or outdated statements (“the browsers have all de-supported SVG”, etc.), I left out of disgust. Fortunately, there was hardly anyone at that talk, so not too many people were corrupted.

I’m sure Ted’s a nice dude, but the talk was simply way off mark.

Comment by Dylan Schiemann — September 27, 2007

Is running the application on the client really more scalable when the client is on battery power?

Comment by Ted Goddard — September 27, 2007

Its interesting that no one has pointed out the real issue. The way I see it there is an inverse relationship between developers utilizing an open, loosely coupled, Ajax toolkit (read: not using JSF) and the size of the yacht this guy can buy…

Comment by Sean — September 27, 2007

Well gee guys I don’t know why you didn’t like his talk. Back end integration really ~is~ the hardest part of making a large scale application work right – if your back end is trying to integrate with some oracle created cluster fuck.

burn them, burn them! seriously…..every organization out there needs to kick these guys(oracle consultants) and all the other vultures/vendors like them to the curb if they want to build a real product someday….

Comment by Jesse Kuhnert — September 27, 2007

…asinine marketing bullshit that has little to do with engineering a web application.

Comment by Mark Holton — September 27, 2007

“the most challenging part – integration with back-end servers.”

Funny. I did it with one page of javascript code and half a page of php code on the server side. Isn’t AJAX popular BECAUSE it is so easy to integrate with the back-end? Isn’t that the point of AJAX?

Comment by Marat Denenberg — September 27, 2007

Just to be clear:

1.) Oracle *buys* a keynote speaking slot at an Ajax conference
2.) Oracle proceeds to diss Ajax, and in particular the wealth of Open Source ajax tools which are available
3.) Oracle pays someone else to develop Ajax components for their JSF “solution”
4.) Oracle plans to Open Source those very components (many of which they didn’t write), thereby adding to the fray

Either this is the worst reporting ever, or Oracle’s sense of irony is non-existent.

Their message here is really strange: be afraid of Ajax unless you buy our product which shields you from it but provides the benefits.

Perhaps they’re just jealous that all their competitors figured out how to work with Open Source projects and contribute in meaningful ways without directly controlling their every move?

Are they serious?

Comment by Alex Russell — September 27, 2007

Look people, it’s the difference between a piece of clientside script to ‘create’ a tree view from data that you get from an endpoint (maybe a webservice, or just a simple GET) and specifying on the server the fact that some data will be a tree view when it’s on the client (like ).

I think we all understand the difference, I use ASP.NET Ajax, so I feel really uncomfortable using something like Ext (will I lose all those ASP.NET ‘web controls’?) and far more comfortable with the Ajax framework that is made especially for ASP.NET (where I get the benefit of page methods, update panels, and so called web control extenders to add client side capabilities in the server side markup).

You need to be familiar with both to understand the true differences, and of course this man has an agenda, and of course the Ext way is not a‘dangerous way to program. But surely, you all must understand the tension between these two? Google for using Ext with ASP.NET (i.e. combining the two), I can’t really find anything. Also, there’s a project to wrap YUI control in ASP.NET control, so you can instantiate them from the server side markup, but even the author acknowledges that integration with ASP.NET is not what it should be.

Very interesting discussion, and really something we should hear more about on Ajaxian. Maybe it’s time for a good editorial?

Comment by Mike — September 27, 2007

4th sentence should read

Comment by Mike — September 27, 2007

With regard to integration with backend servers, I don’t think there are very many AJAX applications that directly connect with back end data stores. Rather, the AJAX approach is generally to develop a web based UI in front of some web based API (e.g. SOAP, JSON, REST, etc). So services are alive and well.

Now, things like Google Gears are starting to change that in some cases, but I don’t see back end services going away. Rather, just as we are starting to see some things settling down on the AJAX UI front (and I use the term settling down loosely and only relative to the chaos of the past year or so), I fully expect that we’ll see things further stabalize on the services front (e.g. a de facto standard for mapping SOAP to/from JSON, a standard approach to securing JSON and REST based service requests via HTTP request headers, etc.).

But JSF is on the way out. There’s no reason for all of that “no man’s land” complexity stuck between the web client and the server when there are much cleaner alternatives like GWT and Wicket for folks who want to use an event based programming model that generates a Web 2.0 UI. And there are many other folks who will decide that having a clean break between an AJAX UI and a set of server side SOAP/JSON/REST/etc based services is a good thing.

Comment by Jim LoVerde — September 27, 2007

I think it should also be pointed out that at least Adobe has a decent solution for talking with backend systems. It’s ColdFusion. While Flex does talk to webservices, it has extra features for communicating with ColdFusion. It can do anything Java as it is Java, and can talk to many backend systems, including but not limited to Oracle.

Comment by Terrence Ryan — September 27, 2007

Having taught Ajax for about 2 years now Ted’s opinion isn’t too off the mark given what I have seen people blurt out – “where’s the server side focus!!!!!” Server-side centric developers are scared out of their wits by a loosely coupled Ajax focused world that treats them as mere data providers to act as an API. Framework vendors certainly play to this and want such folks to develop in their frameworks and not touch that icky JavaScript. Remember for lots of developers JavaScript in their mind is supposed to be a toy language though interestingly it is so easy we won’t bother learning it and just let some component take care of the details. Don’t mind the JavaScript it just makes it work…

These days I introduce the first night of the Ajax class with a comment like “Yes this is a class in JavaScript, advanced JavaScript with remote communications” Interesting reactions then ensue…generally later on we get the admissions about not knowing much about JavaScript and it is kind of cool and what not. The part in a later session where one backend data provider is written in 4 or 5 different frameworks and they just point at a different URL also kind of unhinges a few people. You just aren’t supposed to do that kind of stuff! Come on you are either with us (where us = .NET, Java, Ruby, PHP etc. ) or against us. Isn’t that how we Web developers are supposed to think?

So give poor Ted he doesn’t think that much differently than way too many folks…I say educate them not deride them if you can.

Comment by Thomas Powell — September 27, 2007

JSF is a terrible technology for reasons well discussed on other forums/blogs.

By using JSF one puts the project at risk. If you use JSF in a standard fashion and if you are vigilant and disciplined, you will probably be OK, but there won’t be anything exceptional happening either. On the other hand, should you even slightly step outside the preplanned JSF mind-box, should you do anything even remotely interesting or creative, or should you hit a JSF bug, you are in deep trouble, because then, due to complexity of open-source solutions and due to closed-source nature of other solutions in JSF space, you will have no control over your own destiny.

If something breaks in jQuery, even with jQuery’s fairly ugly source code, I can fix it. I am in control of my destiny. If something breaks in Prototype, I can fix it. I can probably fix it quickly too. Good AJAX Javascript frameworks are transparent and easy to read and maintain. This cannot be said about any JSF implementation. JSF has a very low power-to-complexity ratio. So, I suggest people definitely take what Oracle is saying with a grain of salt.

Comment by Leo Lipelis — September 27, 2007

…yeah, agree with Alex.

…in many areas, Oracle (and MicroShaft for that matter) are great at ‘getting in the way’ of innovation, despite their massive resources. I personally have little use for their products in comparison to innovative open source solutions.

Comment by Mark Holton — September 27, 2007

I say it’s a matter of framework, the current crop things in ajax are just big goto’s. Seaside is the start of an answer I hope more frameworks will look into being more like it.

Comment by sf — September 27, 2007

I actually agree with Ted 100%. He is right. Some of the comments here are ridiculous. I sometimes think developers who come ‘straight into’ Ajax environments (Without having a high end software/application/database background/experience) really just don’t ‘get it’.

@Marat: “Isn’t AJAX popular BECAUSE it is so easy to integrate with the back-end? Isn’t that the point of AJAX?”?

No. No that is not the point of Ajax at all. Not even close. I can’t think of a more invalid statement.

Comment by Gavin — September 27, 2007

Damnit, shanked the closing anchor tag on my last post. Sorry about that guys.

Comment by Alex Shvedoff — September 27, 2007

Just want to remind everybody what AJAX submits look like to the “Backend”… POST and GET/URL encoded variables. Nothing more, nothing less. On the backend, it looks like nothing more than the MC of MVC, without the view (as long as you send data back via json or xml or whatever else you decide). Heck, it could even send back HTML as I do when I get lazy (94% of the time).

Cheers all… I think this topic has been beaten to death with a 10-foot pool. Have a good night, nice hearing the ins and outs of this debate.

Comment by Chad — September 28, 2007

Oh ya, and why is it that we always “debate”? Oh ya, that’s the Socratic method to reaching new levels of intellect, and that’s what we’re doing!!!

Comment by Chad — September 28, 2007

The idea of JavaScript *is* to add nice interface features to an application. Everyone who’s using JavaScript is in the end using a server technology on the back-end to interface with databases. Yes, an interface is extra work, but it’s well worth the effort it takes to build a robust and usable interface on top of a corresponding robust application / database. When you think of JavaScript as an interaction piece, you’re thinking of it in the right way.

Comment by sam parsons — September 28, 2007

Gavin,

I just finished a 2 month project converting a large legacy ASP based website to AJAX in preparation for porting the backend to PHP or JSP, management hasn’t decided yet, heck they might go w/ both! Now instead of having to migrate the previous 200 ASP files I only have 14 ASP files to deal with and of those 14, all of them simply transform a recordset into XML or JSON. This website is now highly portable to multiple web application servers, 100x more scalable, and even easier to update/maintain as the entire presentation layer logic is now contained in 5 standard js files. The old website required 3 servers in a load balanced server farm. After testing we think that we might be able to go back to a much cheaper single server because the server isn’t hardly working at all now under the same number of concurrent users. Business likes AJAX because it saves them $$$$. Developers like AJAX because it is open and easier to maintain and edit.

Sorry for those of you that long to return to the days of the server crunching CPU time to layout a table or having to write a million files because the client always had to bounce back to the server to make a change in the presentation layer. Not even Oracle, Microsoft and Adobe combined can turn back the sea of developers moving towards client based solutions like AJAX.

Comment by Chris — September 28, 2007

Well, I used JSF in some projects and used “web 2.0″ features of JSF. I’m not so excited about it as Mr. Farrell. You don’t need to think about integration with backend, it’s true, but you need to think how to do what you want if this non-standard control or feature. With plain JS and HTML I feel myself freer. For interaction I use RPC, it makes backend and client sides fairly independent.
I thinks a source of this discussion is try to “push JSF”. Yes, it is standard but I do not see this fast growth of JSF popularity which was predicted in the past.

Comment by Siarhei Barysiuk — September 28, 2007

I have worked with JSF few years back, and have written custom components as well. With the advancements in client side JS components and Ajax technologies, I don’t think there is a bright future for JSF or any component architecture at the backend.

Comment by Rajkumar Madhuram — September 28, 2007

Interesting post but even more interesting are these comments.

I’m currently developing a project using JBoss Seam + Richfaces. The presentation tier for this is JSF and although sometimes is does feel heavy compared to some of the PHP dev I have done, overall it is pretty cool and nicely structured.

My major gripe though is the lack of an if/else or switch construct. With JSF you specify if a region on your page is to be ‘rendered’ (true/false). So if you have 2 regions both ‘rendered’ get evaluated.

My philosophy on this is that if you are too rigid and will *only* use JSF on *every* page, then you may find it hard work. If you are flexible and use non-JSF if the page demands it then all will be sweet.

Comment by Damian — September 28, 2007

Have a look at this one. something completely diffrent. it is a paradigm shift for AJAX front ends. One should be aware that I am not, and do not pretend to be objective, never the less I believe that one can judge for himself. Visual WebGui is an open source rapid application development framework for graphic user interfaces of IT web applications. VWG replaces the obsolete paradigms of ASP.NET in both design-time and run-time which were designed for developing sites, with WinForms methodologies, which were designed for developing applications. Thus enabling designer that was designed for application interfaces (WinForms designer) instead of a word documents (ASP.NET designer). This provides the developer with an extremely efficient way to design interfaces using drag and drop instead of hand coding HTML Worth a look at http://www.visualwebgui.com

Comment by navot — September 28, 2007

Guys,

I realize that there are numerous views on the best approach to AJAX. So I will not stoke the fire because as we all well know nearly every solution available has some arguable fault. However, I do want to clear the FUD around some comments made by Alex.

1. Oracle *buys* a keynote speaking slot at an Ajax conference

False. Oracle buys a sponsorship. They were invited to do the keynote irrespective of sponsorship. Thus, if there is an issue with the choice of keynote speaker take it up with Sys-con. It is Sys-con who assembles and approves the content presented at AjaxWorld. This is not the first time Oracle has been a platinum sponsor of the event. Yet, it was the first time the company gave an opening keynote. Moreover, other platinum sponsors did not give opening keynotes much less gold, silver, and diamond sponsors.

2. Oracle proceeds to diss Ajax, and in particular the wealth of Open Source ajax tools which are available

False. Oracle has not “diss”-ed Ajax. The company is an active adopter and contributor to the open source community, specifically in the area of Ajax. Have a look at Apache Trinidad and Apache RCF. They are using the JSF component model to encapsulate the same issues we have all encountered with Ajax. Thus, like everyone else Oracle is trying to simplify development with Ajax. How does that differ from Dojo or GWT for example? Both are trying to simplify and package a complex problem into a reusable solution. You may not agree with the use of JSF – and that is a valid discussion—but like it or not the company is contributing an open source solution under a liberal Apache license.

3. Oracle pays someone else to develop Ajax components for their JSF “solution”

True. Oracle pays the ADF development team –a team comprised of Oracle employees locate in building 200 on Oracle Pkwy in Redwood Shores, CA– to build and maintain these components. The components can be used in an integrated stack, but are available to the general public in their entirety under an Apache License. Moreover, the components are used to build Oracle Applications—a massive suite of enterprise applications. It’s a proven solution and freely available. Anyone is welcome to contribute. If you are have an issue with the solution then you are more than welcome to contribute. I am not sure what more you want here.

4. Oracle plans to Open Source those very components (many of which they didn’t write), thereby adding to the fray

False. Oracle wrote the components. Simply review the list of contributors on the Trinidad project and proposed RCF project to confirm this assertion. Furthermore, Oracle has already open sourced 100+ components with plans to open source another 80+ all of which are in production applications.

In summary, the company is doing a hell of a job trying to do the right thing in terms of contributing to the open source community. Trinidad and RCF are only two examples. You may not agree with JSF as an Ajax solution and that is completely fine. Argue about the technology all you want. However, going on a rant and making wild unfounded claims is unwarranted. If you want to do something productive then help promote more quality contributions from industry to the open source community that are not simply dead drops. Dojo has its own history with IBM as does Eclipse and several Apache projects. All of these projects have had a hand from industry. Does that mean we have to praise all of industry’s efforts? No. But, simply put hearsay and conspiracy theory get us no where. Such claims are no more helpful than pissing in the wind. Let’s try something more informed and constructive.

Comment by Ric Smith — September 28, 2007

Eh… I like Oracle mainly because it isnt MS-SQL.

Well, that and Oracle jobs pay bloody well.

Not that thoughtful of me, but hey.

I think there is a point, but its more one that encapsulates a bit of a stress I’ve had comeing into the Ajax world. In many respects I’m still bit of a “Web 1.0″ sort of guy. I like clean technology, with a server-side focus and nice clean mark up.

Ajax on some level worries me , because its currently a really messy tech. The mix up between javascript and html reminds me on some philosophical level of the god-awful php code we used to, and still oft do, that just mashes table layout html with active code in one file and creates all sorts of head aches trying to maintain and understand whats going on.

But in Ajax’s defense , its still a really young technology, and we’ll see frameworks and tools that beautify the process and make it less brutal to try and make nice things with. We are already seeing them, and *perhaps* JSF is one of those things (Don’t know, havent looked deep into it. Don’t really like java).

One of those wake ups was playing around with the Delphi for PHP thing. Nice idea, still somewhat rusty, and I’m not sure PHP is going to be the thing that delivers us into the new world of really vibrant true web delivered apps. But the potential just hits you in the face. There are plenty of good abstraction layers out there as well that given time really should clean up a lot of the mess. Googles and Yahoos toolkits are very nice, and currently I have a huge crush on Qooxdoo. Still doesnt answer the abstraction questions, but it sure is *pretty*.

Give it time, and ignore the worried suits. But don’t forget the back end. Server issues are neglected at terrible risk.

Comment by Shayne — September 30, 2007

Ric:

Really? You’re gonna go there? Alrighty. Game on.

First, you and I both know the Sys-Con game is a dirty, rigged slime-fest. Keynotes (and other conference talks) are dolled out in large part based on sponsorship. No, it’s not 1:1, but there’s a large enough correlation to clearly suspect causation, even if nothing so crass as out-and-out sales of the speaking slots is ever discussed. Wink wink. Friends of mine who sponsor are routinely “offered the chance to present”, and folks who simply have good technical content but won’t buy a booth don’t get their calls returned. It’s tacky.

As for what Oracle did or did not say (I didn’t go to the talk), all I can tell you is that the discussions of the talk at the conference all seemed to be of the form of “did they really say that?”. You’ll note that I’ve left open the possibility of poor reporting, but given that so many Ajax technology vendors push their products based on fear of “raw Ajax”, it’s not unlikely that this same old tried-and-false trope would be trotted out yet again. I’d love to get a transcript of the talk to put this all to rest, however. Perhaps you can arrange that?

As for who wrote the components, lets not drag Oracle’s contractors into this any more. The good folks you contracted with for a lot of the client-side magic deserve better treatment than that. This is still a very, very small community. You guys should know that by now.

Anyway, I applaud Oracle’s work at Apache. It’s great that you’re being so active in getting stuff opened up. Regardless of whether or not JSF is good technology (it’s pretty clearly not…at least not until the spec gets rev’d at least once more), having good Open Source implementations does a lot to raise the bar for everyone.

Regards

Comment by Alex Russell — September 30, 2007

This discussion is somewhat misdirected — the purpose of Ajax is to improve the user experience by hiding network operations. There are valid ways to do this using server-side techniques (such as JSF) and valid ways to do this using client-side techniques (such as writing application logic in JavaScript). It’s important to choose the technique that your developers and designers will be successful with.

Comment by Ted Goddard — October 1, 2007

I second Ted’s comment. And, thanks Alex for recognizing Oracle’s work with Apache.

Comment by Ric Smith — October 1, 2007

Unlike most of you I was actually at the talk, and I think I have a unique perspective since I also gave the talk. A bunch of this thread is rants on mis-information and things I did not say. It seems like some people never miss an opportunity to tell us how much they hate Oracle. Everyone has the right to their own opinion, but most of the time it is not factual and has nothing to do with the thread that is being talked about. That part is pretty disappointing as it distracts from the people who are trying to take part in a constructive dialog.

My talk was about Web 2.0 in the enterprise. It covered much more than AJAX. w.r.t. AJAX, what I actually said was that AJAX is very cool, but:

a) It is not enough to build full enterprise apps. You need graphing/charting as well as portlet support. Neither of which fall under the Ajax umbrella. This means that you need a solution that is good at integrating a variety of technologies in a consistent manner.

b) I am betting that we are not done with where this technology is going. It will continue to evolve/change and enterprises need to plan for that since they move at a much slower pace than some of the consumer-based sites or smaller niche applications.

c) You should be careful when writing a bunch of client-side code because it makes keeping up with technology changes much more difficult if you have to change thousands of pages (opposed to a single framework) to adopt new or changing technologies.

I also mentioned that there were lots of good frameworks out there, and we (Oracle) decided to go with JSF because it was declarative opposed to code which has a bunch of benefits to enterprise developers over a code-based framework. That was about it. I had a conversation with Gavin who wrote the article sited above and dove in deeper into some of these areas.

So I am always interested in getting constructive feedback and for those of you who did/tried on this thread, thanks. I think some of the things that were brought up here were interesting and things that we spend a bunch of time thinking about as well (eg. perf/scalability with client-side vs. server-side implementations). If any of you would like to continue this thread, please feel free to email me at ted dt farrell att oracle dt com

Ted Farrell
Oracle Corporation

Comment by Ted Farrell — October 1, 2007

Leave a comment

You must be logged in to post a comment.