This content has been marked as final. Show 36 replies
You know what I find truly pathetic? That all these "new & fancy" web GUI features have existed in stateful GUI clients since the mid-90's.
IT IS NOTHING NEW!
But because the web stateless client architecture is such a piece of smelly dung, we now have to marvel that after more than a decade, the stateless web client is achieving what was Common and Standard in GUI design and interaction 10 years ago..
Never mind that in the last 10 years, the GUI bar has been significantly raised and it will take the stateless web client another 10 (or maybe 20 years) to catch up.
Marvelous? Nauseous is a better description in my opinion.
That's a poor framework indeed :P:P ... so poor that Adobe, Cisco, Siemens, IBM, Dow Jones and others more are using it!
I always thought that those companies made nauseous decisions!
I think it's really cool stuff that would be beautiful to have in APEX.
There is jQuery , Dojo , YUI , script.aculo.us. each has different features and different pros and cons.
blog : http://carlback.blogspot.com/
apex examples : http://apex.oracle.com/pls/otn/f?p=11933:5
> That's a poor framework indeed :P:P ...
One that you do not seem to understand?
I'm talking here about business client server architecture - of however many layers you want to throw in the mix. Business applications. The so-called mission critical corporate enterprise systems that requires a flexible, efficient, robust and secure client.
What does a stateless client (web browser) do better than a stateful o/s client in this regard? Answer - sweet nothing at all!
And what is the The Browser that is used by close on (or over) 90% of web users? Microsoft Internet Explorer.
So this whole concept of o/s independence is a failure that the web client-server architecture was suppose to herald is bs. Microsoft guaranteed that.
What is the single biggest open barnyard door in the security of computers today? The Web Browser. Thanks again to Microsoft.
How does the GUI widgets available in the The Browser today compare with the GUI widgets and features of a native o/s application? It does not. The Browser is, as I said, more than a decade behind.
And then there are issues dealing with lack of state, having to scale, optimistic locking, etc. etc.
> so poor that Adobe, Cisco, Siemens, IBM, Dow Jones and others more are using it
Simply because major companies are using that architecture by no means imply that they are using the best architecture.
Just because Microsoft Windows is the most widely used desktop o/s in the world does not mean that is the best o/s in the world. In fact, Microsoft clearly shows that technological superior architecture and design are secondary to how a marketing wars are fought and won.
So implying that the stateless web client-server architecture is superior than a stateful client-server architecture just because company ABC uses it, is just plain silly. (along the same lines as the only Superpower claimed there were WMD's in country X, so it must have been true)
My criticism is not about APEX, or Web2 standards, or web browsers for that matter. I'm not only a fan of APEX, but have over 30 APEX systems deployed in production.
My criticism is HOW the web architecture is used and how poorly it compares against a stateful architecture. NONE of the original reasons (primarily about support and deployment of client software) that swung in favour of web architecture is valid anymore.
But it seems that you, like so many others, have only ever used The Hammer in the software engineering toolbox and thus all problems seem like nails.
Broaden your horizons a bit.
I don't have an axe to grind on either 'side' of this discussion, but I'd be grateful if you could point me to arguments which caused you to reach the conclusion below. Do you equate "web architecture" to SOAP? (whatever that is...)
[[My criticism is HOW the web architecture is used and how poorly it compares against a stateful architecture. NONE of the original reasons (primarily about support and deployment of client software) that swung in favour of web architecture is valid anymore.]]
Thanks very much in anticipation of understanding some of the discussions unfolding in my work place.
> if you could point me to arguments which caused you to reach the conclusion
Over 20 years of client-server development covering most architectures?
> So you equate "web architecture" to SOAP? (whatever that is...)
SOAP is basically just a messaging protocol that uses XML over HTTP(S). But no - that is not what I'm foaming at the mouth about and ranting and raving against.
It a nutshell, I'm sick and tired of the limitations of the web browser - the Thin Client. It is very restrictive ito how you interact with the application and database tiers as it is stateless. It is very restrictive in how the GUI for an application can be designed.
In order to overcome these restrictions, new "technologies" like AJAX, DHMTL, XHTML, etc. etc. have been developed...
Fact - these are still inferior to a stateful client (e.g. A Delphi/VB/C#/Java/C++/Powerbuilder/etc client).
With a proper client, you can fully integrate with the operating system. And that goes by one name only, Windows. Hell, the thin client being used today is Internet Explorer. So any counter argument of o/s independence is not just really silly, but utter bs too (until such time as Microsoft's desktop market share is substantially less than what it is now).
With a proper client, you can do a proper GUI - using all the GUI widgets available on the o/s today to design a meaningful, efficient and good looking GUI.
With a a proper client you can use the local platform resources available. Is it not just plain stupid that after all these years, we have clients that rivals the computing powers of the servers of old.. and yet we put thin clients on them? We place bigger processing and resource burdens than ever before on the back-end platforms and the networks that link them.
And developing a proper client requires less resources and less development as there are less complexities and less moving parts.
So why then the thin-client? It is one of that truly weird things that happened in the late 90's. Frustrations with dealing with stateful client deployments and support these, that had everyone grasping at the first available "solution". And web-based browser technology had all kinds of promises... and boy, did the vendors grasped this opportunity to make and sell new products and technologies.
Instead of client-server becoming simpler, it became increasingly more complex.
Just like RIP (Remote Interface Protocol) and RIP clients in the early 90's. But then who remembers the previous (failed) attempt..?
Anyway, it is like belching against a hurricane. The web architecture emporer is stark naked, but no one cares as too much technology and effort and money and resources have been sunk into this emporer.. it cannot not work. And for all my ranting and raving, I too use web architecture for all my applications... simply as I have no choice at the end of the day. Technology is chosen for me. By end-users. Technology is forced upon me. By vendors.
So please just shuddap and bent over. <sigh>
I think I must agree with you here. Just compare an online spreadsheet with a spreadsheet program on your computer...
But since we seem to have no choice in the matter the original posters question is a valid one.
Not arguing that at all Rene.. simply that I cannot be excited about so-called new technologies for APEX that address the symptoms of the failed web browser client, and not address the actual problems.
And it would seem that same at Oracle shares my basic point of view.
Ted Farrell is an executive in Oracle's development tools division and sets their technical and strategic direction in that regard. He said in a recent article:
"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.
The full article can be read [url http://www.regdeveloper.co.uk/2007/09/25/oracle_ajax_jsf/]here.
Not that I'm anti-AJAX or pro-JSF. In fact, I think that even JSF is trying to fix symptoms of a problematic J2EE architecture.. and not the actual underlaying architectural problem itself.
What irks me is that we have swapped some very frustrating technical issues around supporting/deploying fat clients in the 90's, for an even more complex, more expensive, lessor performing, and even more frustrating thin client architecture.
What drives me beyond frustration is the lack of knowledge amongst developers today to understand these variations of the client-server architectures, the pros, the cons, the issues. Instead they are blindly sucking on the teats of the vendors driving this architecture and technologies because of the their complexity.. as that translates into more h/w, more s/w, more skills sets, more resources - and thus more products and more sales.
Okay.. so maybe I should not complain about vendors doing an excellent job wrt job protection for us software engineers. ;-)
But still.. it is time that today's developers, that seem to know nothing about what client-server architecture is, to wake up to the fact the client-server can be far richer, and indeed a lot simpler.
[[So please just shuddap and bent over. <sigh>]]
Not sure if this is directed at my question, or an acknowledgement of expectation triumphing over hope!
I do agree with you in as much that I think the 'right' way forward is the one which guarantees highest income for longest number of years (for MS and all developers who are touched by its presence).
So I wasn't questioning your credentials, simply asking if there were papers / forums which you think contribute to recent considered discussion of the thick/thin debate. I found your account very forceful, but I'd be fascinated to read of opposing views.
(btw thanks for the above mentioned link)
> Not sure if this is directed at my question, or an acknowledgement of
expectation triumphing over hope!
Nah.. just me spouting off in general. Er.. and please pass the lube. ;-)
> I do agree with you in as much that I think the 'right' way forward is the
one which guarantees highest income for longest number of years (for MS and all
developers who are touched by its presence).
The right way forward in my view is to simply rethink the current 3 tier web client-server architecture, and rip out that which does not work, and replace it with that which do work.
Take the actual wire protocol.. TCP. What a waste on a stateless application protocol like HTTP. SMTP uses TCP, yes. But SMTP is a stateful app protocol. So is POP3. IMAP. NNTP. IRC. FTP.
But some bright genuis decided to run applications over HTTP. Using none of the advantages that TCP provides as HTTP itself cannot care.
IT DOES NOT MAKE SENSE. From the wire protocol, to the application protocols, to the thin client. It simply does not make sense in the 21st century to muck about with what is truly idiotic technological decisions and concepts of the previous century.
But we have big pipes and the most powerful computing platforms ever built.. and that makes this abortion of a technology works and hides The Ugliness Underneath.
Oh dammit.. there I go ranting and raving again. Time for some coffee and reading Dilbert... :-)
Billy: You make some very good points.
Just curious, are you aware of the new (circa 2004 ) "Smart Client" technology? That would seem to address all your pain-points regarding the admittedly inane and frustrating thin-client architecture.
Billy, you're arguing about statefull vs stateless application from only one point of view: technical. All companies out there have only one goal in mind: making money. Technical characteristics are definitely not the first priority when software decisions are made.
I think we all understand you Billy, but the reality is that web applications is a powerful trend and we all must use this ugly stateless client! Of course i love Oracle Forms, Delphi, etc... I developed Oracle Forms apps for about 10 years, and APEX is the most friendly RAD that I can play with to develop web applications. My company is developing big complex applications on APEX... so... I sure have some UI issues that I would love that APEX would take care on future versions!
For instance, yesterday some hours had been "lost" in creating some PPR processes on APEX! So, these new frameworks like EXTJS, Dojo, etc ... are indeed very useful for presenting a friendly UI that can be compared to the old 90's C/S apps.
You can't escape to the browser trend... and remember, C/S is disappearing, 3 tier is the new trend, and the browser it's a vital part of the last tier!
So, instead of ranting and raving :) ... I accepted this as a Inevitability! And sure, I want more from APEX and the great TEAM behind it! And because I want more, I expressed my desire of new AJAX features on APEX 4... Much more than the APEX preview that Carl showed us on his APEX 4 preview!
Best Regards you all,