09/14/08 :: [REST] The Restpolitik [permalink]
I don't think it makes much sense to continue talking about REST. It is clear that the RESTafarians at large have absolutely no idea about what they are talking about and they reject SOAP and WS-* simply because it is unconveniently too sophisticated for them, the limits of their minds and the confines of their use cases. Who ever said that WS-* was designed for browser-to-server interactions? WS-* was originally designed to compete with ebXML in the B2B space and later repurposed by Microsoft and IBM to enable interoperability between .net and Java.
It is also clear that there is a small group of die hard RESTafarians who sees REST being overwhelmingly washed away by Web 2.0 and is desperately trying to keep it alive in the SOA space. These RESTafarians do not care about the level of FUD they create in that space or how many projects they disrupt. They are here to keep REST alive. These people will stop at nothing to "make REST happen".
They simply don't realize that REST was designed with the requirements of Web 1.0 (nothing could be more clear in Roy's thesis) so how could it be still appropriate for Web 2.0? Now don't get me wrong, REST instantiated as the architecture of the Web is an incredible piece of work. If Roy had never spent the time to carefully understand the requirements of Sir Tim Berners-Lee and translate them into a technology such as HTTP, the Web probably would have never existed. But here is the catch, the requirements set forth by Tim have dramatically changed nor they had anything to do with B2B or .Net / Java interoperability. A lot more requirements have appeared today that REST alone cannot address. Don't be fooled by the fact that Resource-orientation is an important and timeless concept, you have to make a clear distinction between abstract concepts and technologies when building a given solution.
So today, there is nothing clearer that the RESTafarians have lost their battle: Roy's REST will never become the programming model of connected systems. This is sad, this is again years of innovation wasted by people that did not understand probably the biggest teaching of Roy (beyond REST and resource-orientation): there is tremendous value for an architect to uncover the architecture style that matches the requirements of the system he is supposed to build rather than stupidly applying his favorite style -or the favorite technology of the moment- to any problem that comes to him or her. No these people, today the RESTafarians, keep thinking and inconsiderably yelling across the industry there is a "magic bullet". Yeah, right.
I would like to conclude my posts on REST with two posts, this one, that is going to talk about the RESTpolitik and a second one that will summarize all the discussions I have had and the points I have made, it will aim to be the final nail on the (other) REST's coffin.
This post is an answer to Stu's response. Ironically, I discovered Stu's response last night since a "resource" cannot notify me it has been updated. Yes, I have to constantly poll it, or build a proprietary notification mechanism from behind. But let's pass on this minor design inconvenience the RESTafarians have long left the ground of any reasonable discussion.
I would like to respond to this point today:
"Now, Steve complains of being "stalked". But who is stalking who? (Roy's) REST has nothing to do with WS-*."
Roy has been quite critical of WS-*, just not as vocal. Your approach to dealing with Steve has showed a complete lack of tact or consideration for respectful debate. I don't think you realize the emotional impact your style of argumentation has on people. At times you're fine, at other times you're completely unacceptable.
I had abandoned the REST debate long ago (2004 or so, I recall). What prompted me to start it again about a year ago was an email that my manager sent about a year ago to me where other people in my company were questioning the validity of a SOA initiative using Tim Bray's infamous post on "The End of SOA". He was asking me how we could respond to this email that was circulating behind our back. He did ask me if this was true. In other words his point was not that we should hide it and defend our choices but his question was, what's the best for our company? So I took on to understand and respond to this post and in the process I discovered the politik behind the (other) REST, because -frankly- there is no engineering behind it, it is only about politik: RESTpolitik.
What I discovered was that there was a group of people who were on a mission and they were no Blues Brothers. It was best illustrated by Pete Lacey explaining that he was hired by the Burton Group to make REST "happen". I discovered a group of people for which everything was justified: It did not matter to them how bogus their arguments could be, it did not matter to them that there was absolutely no evidence that REST could do in the enterprise what it had done to the Web, as long as they could trash SOA and WS-*, they would make room for REST and at this point with Web 2.0 looming, any room was good to take. It did not matter to them that some SOA initiatives across the world could be impacted by such bogus claims. This was just collateral damage to their holly war to cease control of our industry, or at least keep REST alive. It did not matter to them how the bottom line of enterprises is affected by delayed investments in connected systems, as long as they can get their little names in the headlines across the web. These people have simply no scruples and no moral compass.
Why do I say this behavior is toxic to our industry? Let me explain clearly. Say you have a SOA initiative or you are trying to start one. You are going to need some budget to expand or start your initiative. As you may know you have to fight for your budget, nobody hands you money without making sure it is money well spent. What happens when a decision maker reads Tim's post? (there will be plenty of people that make sure he reads this post) He or she says, hum... here is a guy who invented XML, who is a big shot at Sun, and pretty much tells me that the money I am about to spend is a waste. What do I do?
Stu, you know me, I am for unity and balance, after all, I wrote WSPER to explain that the way forward was an architecture where Services, Processes, Events and Resources were ALLTOGETHER the foundation of connected systems. I have tried to talk with this bunch, but this bunch does not talk to you, it has a mission and it will never deviate from it. They look down at you, poor idealist, trying to apply some logic to the discussion and they laugh. No Stu, you see, there is unfortunately only one way to deal with a bunch like this, it is to stand up and inflict pain (as much pain as they inflict on the rest of us trying to do our jobs), to the point where their hollow arguments are shattered. There is no other possible response to the Restpolitik.
Principles are no match for this kind of people, they'll steamroll them in a blink of an eye with no remorse. I have been working on SOA for 10 years, I have participated in many standard working groups at W3C, BPMI, OASIS and OAG. I worked with the best and the worst of the people. I worked with people that were committed to deliver the best they could for the greater good of our industry. I have learned so much from people like John Yunker, Dale Moberg, Conrad Bock, Fred Cummins or Monica Martin. I strongly believe that Roy is one of them. Their mission was not to "make something happen", it was to uncover what our industry needed most and give it away, for nothing in exchange.
I have also seen the worst of people playing their petty power games. I have seen "friends" snapping your role in a standard group before your eyes without second thoughts. I have seen people like Duane Nickull whom you invite to a forming working group run away and create his own competing working group, without even telling you. I have seen incredibly good work like WS-CAF trashed and burned because other people wanted something else to "happen". I have seen big companies trying to stall a working group, for no reason, taking control over the chair role without even any discussion with the group, putting in charge a totally incompetent guy and trying to rot the work of the group. I have seen people steeling a spec and running away with it despite the complains of all the authors and contributors. I have seen people coming unannounced to a meeting claiming they wanted to participate in a working group and the next day leak in the press that they left the group because they were not doing the right things. Most of all, I have heard before that XXX is way too complicated and all we need is YYY. In the end of course YYY looks like XXX (actually worse). So I have been there, done that, I have practiced the Tim Brays and Steve Vinoskis of the world and I came to the conclusion that if you do not inflict pain (legitimately) by showing how weak or bogus their arguments and their case are, they will simply ignore you and continue talking their talk and walking their walk: they are simply not here to deliver anything of value other than to them. Again, the recipe is well known, if nobody stands in the way, this kind of people will continue their Restpolitik without any remorse, using a Restpolitik style is the only way they can win the game, they have no other choice because logic and (engineering) principles are not part of their game.
This is for instance the little game that Assaf Arkin and Howard Smith were playing with BPML. I was of course the idiot trying to apply some logic to the design of BPML and explain that yes business processes have a B2B interface and maybe, just maybe, some users need to pop up here and there to do some work. But hell no, Assaf and Howard were claiming I was just a fool and all the work they were doing was perfectly ok (without any B2B or users and roles semantics). We all know how this ended. I forgot to mention in the list up there, that I have even seen a company, Intalio, running away with a spec written by an organization, BPMI (with paying members), calling the spec something else (a flashy name: WSCI) and get it published with Sun, BEA and SAP and then start over with another BPML spec. What a wonderful world.
What did the industry get in return for letting Assaf and Howard play their little games? Nothing. Yes the BPMS market should be as big as the RDBMS market. Yes BPM could dramatically change the bottom line of enterprises across the world and why did it not happen yet? It has not happened because a handful of people put themselves in charge and imposed their narrow views on the whole industry. There is a word for that: the Restpolitik. The Restpolitik is wrong, it is killing our industry and with it, scores of enterprises who do not have the technologies they need.
What will our industry get in return for letting the RESTafarians get away with yet another instance of pure Restpolitik? nothing but lost opportunities. Just like in the BPM space (which by the way REST does not support a bit, but again, that's just a "detail" for these people).
Can you look at me in the eyes an tell me this is not what RESTafarians are doing? What's all this fuss about about the end of SOA, SOA is a failure and "Death-Star"? Can you imagine how much of a Weapon of Mass Destruction that 10 line post from Tim Bray is? And please don't think I am naive enough to think that he, Steve or others have no idea about what they are doing, they may have no idea of the magnitude of what they are doing and how many people are impacted by their little stunts, but they know exactly what they are doing. How much time is wasted across the world by SOA initiatives because the RESTafarians are trying to create a wedge with the usual "feary/fuddy" argument: it's easy, cheap, no need to learn arcane stuff...? it has even a cool and innocuous name: "webby". How charming.
To specifically respond to this: "I don't think you realize the emotional impact your style of argumentation has on people." Please believe me, this is not my personality, and there is nothing personal behind these words. I am actually paying a heavy price myself for it because I am very conscious that future potential employers might actually see that as a liability rather than an asset. But at the end of the day, the Restafarians are leaving us/me with no other choice because of what they are going after and how they are going after it.
So I sincerely hope you will understand all I am trying to do is to fight the Restpolitik and bring it to a stop while helping fellow architects that have to deal with the consequences of this game day in and day out. Will Steve, Stefan, Tim and others have the courage to step forward and bring a truce to this stupid game, I doubt it, they have chosen the "fuite en avant" strategy so far.