Data is a necessary piece of nearly every software solution (I'm sure we can all think of exceptions, but there is data there too). Web 2.0 via mashups makes data integration from various sources relatively easy. Applying the MVC pattern to Web 2.0 we have view and control provided by Flex, AJAX / JavaScript / wikis / feed readers / browser, etc. and the model JSON or just raw data in syndication format. The formal nuances of data, its structure and relations, etc. are not the focus of mashups but rather pulling out relevant information fragments from various sources into a new whole (aggregation, composition, integration, etc.) creates a new information model, that requires new controllers and new views (unless you can reuse those from elsewhere too, but by and large you are pulling in data, aggregating and defining a new format, contollers and views. That's it. So why do IT departments seem to make it soooooo difficult? Job security and FUD.
Web 2.0 is, and provides, RAD (Rapid Application Development) for the Web. (RAD came out of the client server era, and in many IT shops today the term or concept are never used due to negative connotations and associations of an historic nature). RAD was cool and good back in the day as it empowered less skilled application developers to be more productive...today Web 2.0 technologies allow both developers and non-developers to create powerful, compelling softwware solutions with minimal effort. Minimal effort means a faster time to market as well as lower entry costs, redo, scrap, do over, change all become more cost effective and viable (and faster too).
To see how effective web 2.0 can be for you here are few cool tools that I find valuable. Note, I am not affiliated with these companies, developers or communities in any formal business or even friendship manner; I simply like what they have to offer and I think you will too:
First, for any IT people reading this, you should go have a look at ActiveGrid. I blogged about AG before, and will do so again in the future. If I was in charge of a small IT shop (or a team in a large IT organization) that needed to provided higher value, reduced costs and enable greater agility, this is where I would start. I'd build and deploy new apps on AG, retool my IT employees and start a migration plan for existing applications. I'd also start with the open source version until I was nearly ready to go to production and would give careful consideration to a commercial license based on the information available to me. ActiveGrid would be a foundational technology platform upon which I would rebuild my IT infrastructure. Period, end of story.
These other products and technologies could be leveraged by IT too, but are more focused on business people that need to get business done rather than IT who are in the business of supporting business, i.e., they are more usable for non-IT business users.
RSSBus. Just go and have a look, they do a great job of splainin' (explaining) what they do. Basically, you take data from all kinds of sources, aggregate it and create new scripts to act on it, and you can do this recursivley; building up your own catalog of feeds and scripts and using those provided by others. Of course, you are encouraged to submit yours for use by others. Great way to pull all the data you need in one place with a very easy toolset. Kudos to the RSSBus team! I will have more to say about this technology in the future too.
You can combine RSSBus and ActiveGrid in many complimentary ways.
Another technology I wanted to mention is OpenKapow. They claim "mashups in minutes" and its very accurate and true.
There are also Yahoo Pipes and Google Gadgets.
There are some overlaps in these and complimentary combinations too. Please let me know if you are using any of these and where, or if you are using other Web 2.0 tools, I'd like to learn about what else is out there. These are the tools that I would look to use if I was in charge of IT direction at any company. I'm sure there are many others...what they all have in common is simplicity and utilization of core technologies available to everyone on the Web.
Now have a look at what some IT industry heavyweights are proposing for business application development: big, complex, costly and heavyweight solutions that are so last century. I noticed that many of the products of these large vendors are consistent with their own size and complexity, the bigger the vendor the more complex their solution strategy it seems. There is an easier, faster and cheaper way to provide technology solutions to business, its Web 2.0. So easy its almost magic ;)
Thanks for reading
11 May 2007
30 April 2007
The Complexity Proponents Run IT
Its been a busy two months since last I blogged here. I've been working with several Fortune 100 companies as usual (yes well within the F 1000) and the complexity mantra was reinforced again. Why is it when faced with complexity the response is often more complexity? Complexity to fight complexity?! I confess that I don't get it, there may be some organizational psychology or sociology that can explain it, to me its just plain idiocy.
The challenges and problems facing large IT shops are difficult, varied and complex but not insurmountable, unless you don't have an organization equiped to handle it. Unfortunately, as is often the case, most organizations are not prepared to change their thoughts and approach. They will choose complexity. Adding more compexity to meet these issues is not the solution. Most IT solutions are needlessly complex, insufficient for their purpose, difficult to change, manage: overengineered. Instead of overengineering, design in cost effective change enablement and empower end users. (this will be a future blog topic I'm sure!)
There are simple approaches and technologiess that can help, but those aren't used. Instead, the enterprise IT tool of choice is SOA. Complex, standards bound, political, expensive...you name it, it is complexity to solve complexity. Now, I think SOA is good, valid approach however in the hands of the typical enterprise IT shop it quickly gets mired in politics and complexity. Heterogenous IT shops with their various factions, will dilute, fragment and fracture SOA to the point where it will fail to deliver. In practice, SOA's core simplicity is lost, and its complexity becomes the sustaining driver. Each iteration within the political and architectural circles of the enterprise makes the situation worse.
Meanwhile, there's Web 2.0 and simplicity available. It is so comparatively easy in fact, I think that is why it isn't going to be adopted by many IT shops. If its that easy, who needs all these cost-center employees then? IT departments want to simplify but they can't actually bring themselves to actually do it. If and when an IT shop actually does apply easy-to-use technologies successfully, they will be in a position to out manuevre their competition.
Where do you start? Get a Web 2.0 team, from outside, and have them perform a bake-off with your IT department using their tools and technologies. When the Web 2.0 team hits it out of the park, ask your IT department what their problems are, and listen for the excuses. None are valid, it is purely self serving, self protecting doubletalk. They used complexity to address complexity, and that approach failed outright, cost more and underdelivered or some other partial success story.
Is Web 2.0 perfect? Absolutely not, but its premise is based on simplicity and high value. Used with sound (not old school complexity) architectural principles and design aesthetics that focus on simplicity of form and function, Web 2.0 will be better, cheaper and faster than anything the compexity proponents can provide.
How well did your organization adapt to and adopt Web technology (aka Web 1.0)? How about CORBA and/or DCOM? OO? If you didn't get the value you expected from these, you will likely not get the value from SOA you anticipate. What to do?
Go Web 2.0, of course! You need the right team (beware, they could be complexity mavens in disguise!), that can effectivel leverage the tools and technologies that are available in a blend of art, science and engineering simplicity and economy.
Well, what's to worry? Your organization won't do it (no real incentive) and your competitors won't do it either, so you'll all use the status quo vendors and technologies, employ high-cost complex technologies and IT experts and complain or worse offshore to lower the costs (and increase complexity again). Remember:
"Insanity: doing the same thing over and over again and expecting different results."
- Albert Einstein
So, is your IT approach insane? If you are a large enterprise IT shop, the answer is probably Yes!
The challenges and problems facing large IT shops are difficult, varied and complex but not insurmountable, unless you don't have an organization equiped to handle it. Unfortunately, as is often the case, most organizations are not prepared to change their thoughts and approach. They will choose complexity. Adding more compexity to meet these issues is not the solution. Most IT solutions are needlessly complex, insufficient for their purpose, difficult to change, manage: overengineered. Instead of overengineering, design in cost effective change enablement and empower end users. (this will be a future blog topic I'm sure!)
There are simple approaches and technologiess that can help, but those aren't used. Instead, the enterprise IT tool of choice is SOA. Complex, standards bound, political, expensive...you name it, it is complexity to solve complexity. Now, I think SOA is good, valid approach however in the hands of the typical enterprise IT shop it quickly gets mired in politics and complexity. Heterogenous IT shops with their various factions, will dilute, fragment and fracture SOA to the point where it will fail to deliver. In practice, SOA's core simplicity is lost, and its complexity becomes the sustaining driver. Each iteration within the political and architectural circles of the enterprise makes the situation worse.
Meanwhile, there's Web 2.0 and simplicity available. It is so comparatively easy in fact, I think that is why it isn't going to be adopted by many IT shops. If its that easy, who needs all these cost-center employees then? IT departments want to simplify but they can't actually bring themselves to actually do it. If and when an IT shop actually does apply easy-to-use technologies successfully, they will be in a position to out manuevre their competition.
Where do you start? Get a Web 2.0 team, from outside, and have them perform a bake-off with your IT department using their tools and technologies. When the Web 2.0 team hits it out of the park, ask your IT department what their problems are, and listen for the excuses. None are valid, it is purely self serving, self protecting doubletalk. They used complexity to address complexity, and that approach failed outright, cost more and underdelivered or some other partial success story.
Is Web 2.0 perfect? Absolutely not, but its premise is based on simplicity and high value. Used with sound (not old school complexity) architectural principles and design aesthetics that focus on simplicity of form and function, Web 2.0 will be better, cheaper and faster than anything the compexity proponents can provide.
How well did your organization adapt to and adopt Web technology (aka Web 1.0)? How about CORBA and/or DCOM? OO? If you didn't get the value you expected from these, you will likely not get the value from SOA you anticipate. What to do?
Go Web 2.0, of course! You need the right team (beware, they could be complexity mavens in disguise!), that can effectivel leverage the tools and technologies that are available in a blend of art, science and engineering simplicity and economy.
Well, what's to worry? Your organization won't do it (no real incentive) and your competitors won't do it either, so you'll all use the status quo vendors and technologies, employ high-cost complex technologies and IT experts and complain or worse offshore to lower the costs (and increase complexity again). Remember:
"Insanity: doing the same thing over and over again and expecting different results."
- Albert Einstein
So, is your IT approach insane? If you are a large enterprise IT shop, the answer is probably Yes!
07 February 2007
REBOL 3.0
It looks like REBOL 3.0 will be here soon!
First, there is a high-level component architecture overview document here...
... a list of 2007 priorities here...
...and if you follow the link to the REBOL Developer Conference, the program states "Introducing REBOL 3.0" by Carl Sasserath.
Can we hope to see REBOL 3.0 released and available this summer? I can't wait!
First, there is a high-level component architecture overview document here...
... a list of 2007 priorities here...
...and if you follow the link to the REBOL Developer Conference, the program states "Introducing REBOL 3.0" by Carl Sasserath.
Can we hope to see REBOL 3.0 released and available this summer? I can't wait!
Good for you, not for me
Now I'm a big Steve Jobs and Apple fan, but I'm not a fan of something that appears to be hypocritical. I ceratainly don't like "Hollywood" and the rest of the entertainment industry at all, but that 's outside this topic.
Steve's recent open letter on music calling for copyright holder to eliminate DRM is interesting. So Steve, the record companies should eliminate DRM and allow music to be shared and copied however consumers wish? How about Apple doing that with Mac OS X Leopard, let consumers install your OS on any machine they want and as many machines as they want, or even share their copy of your OS with their friends and families? Sounds like "do as I say, not as I do".
I have a few Mac's, can I go and buy one copy of Leopard and install it on all of my Macs? I have a few Intel boxes running Linux, can I install Leopard on them too, using my one copy? I have friends and family that would also like to run Leopard, should I let them use my single copy that I paid for? Of course not, I'm not allowed to do that. Apple would see that as a breach of contract or violating their ownership rights or something of that sort, that is, it is not legal for me to do so. In fact, even if I purchase a separate copy for each machine I own, it is illegal for me to install Mac OS X on a non-Apple machine.
Tell us why should any digital content owner (movie studios, music labels and other publishers, etc.) allow someone to buy their content once and copy it anywhere and everywhere, to any / as may devices (e.g., computers, mp3 players, phones, PDAs, etc.) and share it with anyone they wish?
When Apple relaxes or eliminates its software digital rights, its licensing and usage requirements and constraints on its software (not just the OS), then they can take the "moral" high-ground position and ask others to do the same, not before.
So Steve, when do you plan to allow consumers unlimited copy and unlimited use digital rights for each single copy of Apple software that they purchase? When will consumers be allowed to install Mac OS X on non-Apple hardware?
Steve's recent open letter on music calling for copyright holder to eliminate DRM is interesting. So Steve, the record companies should eliminate DRM and allow music to be shared and copied however consumers wish? How about Apple doing that with Mac OS X Leopard, let consumers install your OS on any machine they want and as many machines as they want, or even share their copy of your OS with their friends and families? Sounds like "do as I say, not as I do".
I have a few Mac's, can I go and buy one copy of Leopard and install it on all of my Macs? I have a few Intel boxes running Linux, can I install Leopard on them too, using my one copy? I have friends and family that would also like to run Leopard, should I let them use my single copy that I paid for? Of course not, I'm not allowed to do that. Apple would see that as a breach of contract or violating their ownership rights or something of that sort, that is, it is not legal for me to do so. In fact, even if I purchase a separate copy for each machine I own, it is illegal for me to install Mac OS X on a non-Apple machine.
Tell us why should any digital content owner (movie studios, music labels and other publishers, etc.) allow someone to buy their content once and copy it anywhere and everywhere, to any / as may devices (e.g., computers, mp3 players, phones, PDAs, etc.) and share it with anyone they wish?
When Apple relaxes or eliminates its software digital rights, its licensing and usage requirements and constraints on its software (not just the OS), then they can take the "moral" high-ground position and ask others to do the same, not before.
So Steve, when do you plan to allow consumers unlimited copy and unlimited use digital rights for each single copy of Apple software that they purchase? When will consumers be allowed to install Mac OS X on non-Apple hardware?
02 February 2007
Enterprise Open Source
I've been collecting an ever growing list of open source offerings that have applicability to enterprise IT. I am convinced that with the right guidance, most enterprise IT shops can do at least ~80% of their IT with these open source offerings. While many of these fall short when compared to their commercial counterparts in terms of features, they still offer a viable alternative for IT organizations with the requisite in-house skills (real programmers; real software engineers).
Let me define "the right guidance". That is someone or some team that can see the value of the Web 2.0 approach to software creation and utilization. A former colleague of mine used to say. "a fool with a tool is still a fool" - if you don't "get" the premise of Web 2.0, if you don't "think" in terms of Web 2.0 or "believe" that Web 2.0 works, then why would it work? The peer-to-peer approach offered a similar opportunity, unfortunately the underlying technology model of p2p was mistaken for the high-profile illegal use of the technology to steal copyrighted content. I worked in p2p for a few years, and was dismayed by the number of people in IT that said, "P2P, isn't that illegal?" What!? (This is similar to the arguments against private gun ownership, "criminals use guns to commit crime, then we must ban guns" Duh!? No incarcerate the criminal is the answer to the actual problem. Remember, "Guns don't kill people, people kill people"!!!)
The feature set is one area, perhaps the most convenient argument open source opponents use; its right up there with the legal / licensing / indemnification "concerns". There are open source integrators that will build, certify and often indemnify an open source suite. So what is the real issue?
The real issue is that it will destroy the kingdom of IT, large numbers of top-heavy management, architects and adiministration overhead (not sys admins, they provide actual value to an organization). If you can find, download, learn, extend and deploy open source solutions to service your customers, employees, partners, execs...why not do it? Using open source will allow organizations to downsize their IT especially in the "deadweight" category. All the money saved can be used to hire real software engineeers and real programmers (there is a difference, which I may blog about eventually).
In applying open source, Web 2.0 solutions in the enterprise (Enterprise 2.0) you will need a few knowledgable business analysts, a small number of tech savvy managers that still have to write or test software themselves, software engineers, real programmers and system adminstrators. Essentially, that's it! For senior management: a CTO and small staff of software engineers that can assess technologies and trends. There are many tools emerging that will enable the end users to have more control over the software systems they use to perform their jobs, they won't need IT to do "requirements analysis" and "BUFD" with heavyweight technologies. Notice there is no CIO position, they are extinct.
If you have a typical, traditional enterprise IT shop of say 100 staff total, I would be surprised to learn that more than 30 actually did anything to create real value (real value: investigate technology, assess needs, write code, test code, admin runitme environments), let's say its 50 or 60 out of that 100 total. Eliminate 40 non-value producing positions, maybe create 5-10 new value producing positions (CTO and advisory staff) and provide better end user experience, cheaper and faster?! What's that worth? You flatten the IT orgainzation with fewer layers of political indirection, make IT more agile and more supportive of your organization's business objectives.
One caveat, having real software engineers and real programmers you will need to have technically competent, motivated leadership (CTO and staff) to harness and focus the creative energy. Clearly defined business objectives, ample time for creative exploration (10-20%) and effective n-ary collaboration and communication is all that is required. (That and good, fast, machines from any manufacturer they want, with big/multiple monitors, a break room with tv, pool (swimming or table!), pingpong, foozball, air hockey, and good coffee, tea, softdrinks etc.) You'll still save money right off the bat, and on an ongoing basis.
There is a better way to do Enterprise IT, for those willing to "risk" it. You can start small and evolve, it doesn't have to be a violent overthrow though the start small team will have many enemies circling like vultures, so executive management (CEO and staff, minus the CIO) will need to protect this team. Those organizations that follow this path will win competitively. Google, MySpace, YouTube are but a few examples of how to do it the Web 2.0 way. What's stopping you?
Let me define "the right guidance". That is someone or some team that can see the value of the Web 2.0 approach to software creation and utilization. A former colleague of mine used to say. "a fool with a tool is still a fool" - if you don't "get" the premise of Web 2.0, if you don't "think" in terms of Web 2.0 or "believe" that Web 2.0 works, then why would it work? The peer-to-peer approach offered a similar opportunity, unfortunately the underlying technology model of p2p was mistaken for the high-profile illegal use of the technology to steal copyrighted content. I worked in p2p for a few years, and was dismayed by the number of people in IT that said, "P2P, isn't that illegal?" What!? (This is similar to the arguments against private gun ownership, "criminals use guns to commit crime, then we must ban guns" Duh!? No incarcerate the criminal is the answer to the actual problem. Remember, "Guns don't kill people, people kill people"!!!)
The feature set is one area, perhaps the most convenient argument open source opponents use; its right up there with the legal / licensing / indemnification "concerns". There are open source integrators that will build, certify and often indemnify an open source suite. So what is the real issue?
The real issue is that it will destroy the kingdom of IT, large numbers of top-heavy management, architects and adiministration overhead (not sys admins, they provide actual value to an organization). If you can find, download, learn, extend and deploy open source solutions to service your customers, employees, partners, execs...why not do it? Using open source will allow organizations to downsize their IT especially in the "deadweight" category. All the money saved can be used to hire real software engineeers and real programmers (there is a difference, which I may blog about eventually).
In applying open source, Web 2.0 solutions in the enterprise (Enterprise 2.0) you will need a few knowledgable business analysts, a small number of tech savvy managers that still have to write or test software themselves, software engineers, real programmers and system adminstrators. Essentially, that's it! For senior management: a CTO and small staff of software engineers that can assess technologies and trends. There are many tools emerging that will enable the end users to have more control over the software systems they use to perform their jobs, they won't need IT to do "requirements analysis" and "BUFD" with heavyweight technologies. Notice there is no CIO position, they are extinct.
If you have a typical, traditional enterprise IT shop of say 100 staff total, I would be surprised to learn that more than 30 actually did anything to create real value (real value: investigate technology, assess needs, write code, test code, admin runitme environments), let's say its 50 or 60 out of that 100 total. Eliminate 40 non-value producing positions, maybe create 5-10 new value producing positions (CTO and advisory staff) and provide better end user experience, cheaper and faster?! What's that worth? You flatten the IT orgainzation with fewer layers of political indirection, make IT more agile and more supportive of your organization's business objectives.
One caveat, having real software engineers and real programmers you will need to have technically competent, motivated leadership (CTO and staff) to harness and focus the creative energy. Clearly defined business objectives, ample time for creative exploration (10-20%) and effective n-ary collaboration and communication is all that is required. (That and good, fast, machines from any manufacturer they want, with big/multiple monitors, a break room with tv, pool (swimming or table!), pingpong, foozball, air hockey, and good coffee, tea, softdrinks etc.) You'll still save money right off the bat, and on an ongoing basis.
There is a better way to do Enterprise IT, for those willing to "risk" it. You can start small and evolve, it doesn't have to be a violent overthrow though the start small team will have many enemies circling like vultures, so executive management (CEO and staff, minus the CIO) will need to protect this team. Those organizations that follow this path will win competitively. Google, MySpace, YouTube are but a few examples of how to do it the Web 2.0 way. What's stopping you?
29 January 2007
Take a REST from the SOA standards insanity
Every day it seems, I become more RESTful. As I look out over the ever increasing morass that are the SOA / web service standards, emanating from a plethora of standards bodies, consortia and vendor alliances I often wonder, when the complexity mavens will cease complexity expansion. Unlike global warming, this complexity is man-made.
Having worked in and continue to work in SOA, EAI, JavaEE, CORBA and other distributed solutions and technologies for many years, thre is one thing these all have in common: All have suffered from design-by-political-committee aspects and/or complexity which limited their usage and market growth potential. What has the IT industry learned from these efforts?
Well, I learned and experienced and still live in this morass filled enterprise SOA world. Unlike some, after I learn and experience, I like to improve, I like to simplifiy what I have learned and experienced, so I thought: how may one make sense of all of these past attempts, from all these growing numbers of standards and technologies for web computing for business, government, science and more? How may one separate the necessary from the unnecessary? How can one do SOA in a simplified manner? What are the basic elements required to enable SOA, b2b, m2m, s2s, c2b, c2c, etc., in a scalable, fault-tolerant, manageable manner with dynamic discovery, production and consumption? What is a viable, basic, simple model? One doesn't have to look too hard to see examples of a simple successful model for SOA web services, but one needs to know the real / actual from the mere usurpers of the term who often live in marketing... the simplification? Web 2.0.
I won't provide definitions for SOA, Web 2.0, REST, etc., those are all over the web. What I want to do is look at where SOA/Web Services emerged and where REST / Web 2.0 emerged, just a comparison that I hope will show the utter futility of one approach that is destined for limited success like many of its enterprise predecessors until it collapses under its own weight and complexity vs. the simple, lightweight alternatives that are providing value now, at much lower cost of ownership, and continue to grow their value proposition exponentially! This is just a stream of thought. You will notice that I am a firmly convinced proponent of REST and Web 2.0 approaches, though I still work in a world dominated by SOA web services and the never-ending standards explosion, so here I vent:
Let's start by abstracting and describing the basic problem model. First, let's divide the web world into two roles: Consumers and Produers. Next we'll divide this world into two perspective or approach camps: Simple and Complex. For each role, either perspective or a combination, may be used. This combination is happening on rare occassions, invariably though the complexity takes over.
Consumers and Producers
The web world is divided into producers and consumers, though these roles may be, and are often, dynamic and reversible, depending on the context, i.e., a producer may also consume and a consumer may produce information in a given interaction. For example, when one uses a search engine one enters search criteria which is consumed by the search engine, which produces results to the requester (the original consumer) who consumes the response.The search engine consumed the request however the search engine may use the original search request and subsequent selections to do more than simply create a search response, it is, in effect, consuming the request which will assist its production / servicing of future requests. By providing value to consumers, the service provider also receives value from those who consume its services.
Producers produce content or information that some consumers may want to consume. So producers need to decide what they want to produce and how to let potential consumers find and consume what they produce. Producers typically produce information with a goal that their information be consumed so they must find a way for potential consumers to consume the information they produce.
As a producer, or service provider, I need to let potential consumers find me, to learn about what information I produce and what I require in order to produce the information. For example, if a producer offers to provide a stock quote, it should tell potential consumers that fact, as well as what consumers must do to get a quote and what that quote response will look like in form, content... Provide a valid stock symbol perhaps, a currency and maybe a stock exchange? What are the defaults and acceptable values? Does the Consumer get back a dollar amount? Are these delayed or real-time quotes? Are there charges or disclaimers, etc....
Consumers need to be able to find producers that produce the information that they are interested in consuming, understand the requirements for consuming a producer service, expectations and the terms and conditions of use, etc. and the format(s) for the request and response supported, required and offered by the producer.
So far this is simple, there are plenty of web sites that provide static and/or dynamic content, aggregator web sites, web services that provide content / information and mashups that combine and consume the products of multiple producers to provide a new whole...then the complexity can start in, if one is so disposed to allow it.
Sometimes the interactions between consumers and producers have formal definitions, possibly one or more alternatives. For example, a producer may provide a URI to which one may simply send an HTTP GET, or have a description that describes how to format an HTTP POST request or maybe it is an RSS or ATOM feed at this URI, maybe its more formalized....you may have a set of parameters and values, or an XML document described by XSD and/or where you need to authenticate via HTTPS or maybe its a SOAP endpoint whose address is defined using a WS-Addressing endpoint reference, whose operations and requirements are described in WSDL, according to WS-RF or WS-Transfer or both via WS-Resource Transfer (are you with me on the growing complexity?!) with a formalized SLA defined by some other standards, with transactions defined by other standards, with parameters and formats defined by several schemas and namespaces and resources and .... and.... and...Whew!
Now, I'm all for consistency and clearly defined descriptions, definitions, formats, etc. but at some point all this complexity adds up: adds cost, reduces performance, increases error potential, becomes more difficult to use, etc. all of which overshadows the whole reason why a consumer may have wanted to consume the producer's product in the first place....to get something of value done. That value may be profit, learning, etc., but it is useful work. If I have to put a dollar in and my ROI is twenty cents, why would I do that? That's like making charitable contributions to the Red Cross where much of what you donate goes to pay top executives, not to the people who actually need it, you know the one's you're making the donation to help!
So, why work so hard and get back less value than you put in? If you go down the standards-based SOA web service path...well, there you are and there goes your ROI!
If you want more consumers to consume your produced products, you need to provide access as easily and cheaply as possible... so that it is consumable by more consumers. it goes without saying that your information product must be accurate, correct, precise and valid! If your producer is not easy to use then someone else will provide a product that is better, cheaper, faster and/or easier and get used by more consumers....if a consumer needs a CS PhD to use your producer, to consume your content, well there are fewer CS PhD's out there than entrepreneurs, high school students, Moms, Dads, Little League teams, and small businesses and other types of information consumers...you need to make it easier for anyone/everyone...so 3 simple rules
Rule #1: Simple is Better than Comlex. Always.
Rule #2: Simple does not imply less rigor, just less work i.e., work smarter not harder.
Rule #3: If its getting more difficult or complex you are probably doing something wrong (or you have a CS degree and just can't help yourself. I have a CS degree and I often need to remind myself of these rules!)
Perspectives: Simplicity vs. Complexity
Survey the Web. Look at all the most successful information producer / consumer / mashups out there and see what they are using, what their technologies are, where the information they consume comes from and the form of the produced/consumed information. Now, look at what enterprise IT shops are doing, just do a search on jobs for “web services, SOA, architect, enterprise” or look at IT job posting for the Fortune 1000 companies...better yet go read some specs at oasis now put these specs to good use in a simple, easy to use (produce/consume) solution.
What you will find are the two basic camp mentioned before, that I call: 1) the simplicity camp and 2) the complexity camp.
First, I'll describe the complexity camp: Many in the complexity camp work in enterpriise IT. The complexity camp is exemplified by / are big companies with big IT budgets and staff that do much processing over the Web, both internally within their organization and externally with customers, partners and suppliers. They have a lot of legacy systems made up of all those prior technology efforts that linger on, seemingly forever...there is never time to do more than keep them on life support so they in turn continue to provide a source of complexity that grows over time. The complexity camp likes standards. It likes standards so much, that it wants lots of 'em. Closed, old, new, contradictory, overlapping, vague, constrainging and open standards. All these standards then require complex products that are used to create solutions using all these myriad standards. Many of these standards have alternatives from other standards bodies that do some of the same and some different things, so you need people on staff to help you navigate all the different standards, their status and their future direction and to provide coordination of all these standards, etc. Once you have that figured out, you'll need to add your own corporate, division and area standards and interpretations, maybe some SixSigma or other fifth element processes and procedures, and of course more staff to oversee this too. Next, you'll need to have some heavyweight tools to develop solutions using all these standards too. Yes, there are some open source tools and technologies you can use, but to be most efficient you'll probably need to buy some commercial tools too, and these will cost you time and money, maybe small money, maybe big money...it all depends...on what you want to do, and how, and why and how frequently, and other factors. Then of course, you'll need weeks of education to learn how to use those tools. Finally, you are prepared to develop your customer solution. What were those requirements again? So the complexity camp is often found in traditional, enterprise IT staffed with the alchemists of the BS in CS variety where: Complexity is enterprise IT's rationale of existence; just business as usual.
Ask any CEO or business executive if they would be interested in reducing the overhead, costs, size, etc. and/or improve the responsiveness, ROI and effectiveness of IT. Care to guess what their answer may be? And so many people in IT wonder why IT is being outsourced and off shored? Hello! Hello? Anybody home?
Like so many human organizations, that become self-fulfilling and self-sustaining, carrying on the work, form and vestiges of the past long after it ceases to be required... non-profits to cure incurable diseases, civil rights organizations that now advocate for preferential treatment rather than actual equality, etc., Not all, but many IT shops are like this too. Most enterprise CIO's and IT shops are like dinosaurs that haven't realized they are extinct! Case in point, what enterprise IT shop has had to deal with the growth rates and capaciaty expansion experienced by inventors / practitioners of Web 2.0 concepts such as Google, YouTube and MySpace? Few if any! Nor could their IT shops assimliate such growth / speed if they had to! They would collapse! Case Closed.
I make no attempt to hide where my sympathies are, just look at the reality of the Web today, who are the players, the movers, the shakers and the leaders? Web 2.0 companies! The Simplicity Camp is the only realistic, practical and pragmatic way to go if you want to provide viable solutions in a cost effective manner.
The Simplicity Camp
Tim O'Reily initially defined Web 2.0 and many others have adopted and expanded the concept, much is being written about it and more importantly, much is being done with it now, today!
Driven initially by constraints (money, time, other resources, etc.), the simplicity camp looked for alternative ways to approach the problem of web computing, doing more with less. Tried and true technologies of the Web such as HTTP, HTML then XML*, XHTML, etc., low/no cost such as open source software stacks such as LAMP, MAMP, SAMP or even WIMP, then more open source emerged that could assist the simplicity camp, simple more efficient technologies and approaches often based on the tried and true technologies of the Web: HTTP, URI, XML, Namespaces --> and a seminal dissertation by Roy Fielding ---> REST and with the advent of WEB 2.0 a new, more simple way was defined.
The emergence of complementary technologies such as RSS/Atom, Wiki's and AJAX, and producer services available from some big companies in a simple, consumable format i.e., one did not need to have a CS degree and arcane alchemic knowledge of SOA web services, i.e., SOAP, WSDL, UDDI, and more WS-* specs, standards in order to consume these offered services from Google, Amazon, Yahoo and many other well-known industry giants. In fact, more consumers consume information from producers using plain old HTTP GET/POST, REST and/or Web 2.0 than via the standards based, enterprise SOA Web Services. Recently, I heard Google was "sunsetting" = phasing out, its SOAP API for search in favor of an AJAX API (AJAX is a cornerstone Web 2.0 technology which has been so successful that commercial and open source vendors of web services based SOA products have added it to their products! Wonder what that really says?! )
While the standards morass around web services based SOA plods along meandering down its many political paths, the REST / Web 2.0 proponents are providing simple, elegant solutions that enable the web for all to consume, produce and more.
If you're in an entrprise IT department doing SOA and Web Services, and "Enterprise" and/or "Architect" is in your title, you need to have a look around outside your walled garden.
Which camp are you in?
Having worked in and continue to work in SOA, EAI, JavaEE, CORBA and other distributed solutions and technologies for many years, thre is one thing these all have in common: All have suffered from design-by-political-committee aspects and/or complexity which limited their usage and market growth potential. What has the IT industry learned from these efforts?
Well, I learned and experienced and still live in this morass filled enterprise SOA world. Unlike some, after I learn and experience, I like to improve, I like to simplifiy what I have learned and experienced, so I thought: how may one make sense of all of these past attempts, from all these growing numbers of standards and technologies for web computing for business, government, science and more? How may one separate the necessary from the unnecessary? How can one do SOA in a simplified manner? What are the basic elements required to enable SOA, b2b, m2m, s2s, c2b, c2c, etc., in a scalable, fault-tolerant, manageable manner with dynamic discovery, production and consumption? What is a viable, basic, simple model? One doesn't have to look too hard to see examples of a simple successful model for SOA web services, but one needs to know the real / actual from the mere usurpers of the term who often live in marketing... the simplification? Web 2.0.
I won't provide definitions for SOA, Web 2.0, REST, etc., those are all over the web. What I want to do is look at where SOA/Web Services emerged and where REST / Web 2.0 emerged, just a comparison that I hope will show the utter futility of one approach that is destined for limited success like many of its enterprise predecessors until it collapses under its own weight and complexity vs. the simple, lightweight alternatives that are providing value now, at much lower cost of ownership, and continue to grow their value proposition exponentially! This is just a stream of thought. You will notice that I am a firmly convinced proponent of REST and Web 2.0 approaches, though I still work in a world dominated by SOA web services and the never-ending standards explosion, so here I vent:
Let's start by abstracting and describing the basic problem model. First, let's divide the web world into two roles: Consumers and Produers. Next we'll divide this world into two perspective or approach camps: Simple and Complex. For each role, either perspective or a combination, may be used. This combination is happening on rare occassions, invariably though the complexity takes over.
Consumers and Producers
The web world is divided into producers and consumers, though these roles may be, and are often, dynamic and reversible, depending on the context, i.e., a producer may also consume and a consumer may produce information in a given interaction. For example, when one uses a search engine one enters search criteria which is consumed by the search engine, which produces results to the requester (the original consumer) who consumes the response.The search engine consumed the request however the search engine may use the original search request and subsequent selections to do more than simply create a search response, it is, in effect, consuming the request which will assist its production / servicing of future requests. By providing value to consumers, the service provider also receives value from those who consume its services.
Producers produce content or information that some consumers may want to consume. So producers need to decide what they want to produce and how to let potential consumers find and consume what they produce. Producers typically produce information with a goal that their information be consumed so they must find a way for potential consumers to consume the information they produce.
As a producer, or service provider, I need to let potential consumers find me, to learn about what information I produce and what I require in order to produce the information. For example, if a producer offers to provide a stock quote, it should tell potential consumers that fact, as well as what consumers must do to get a quote and what that quote response will look like in form, content... Provide a valid stock symbol perhaps, a currency and maybe a stock exchange? What are the defaults and acceptable values? Does the Consumer get back a dollar amount? Are these delayed or real-time quotes? Are there charges or disclaimers, etc....
Consumers need to be able to find producers that produce the information that they are interested in consuming, understand the requirements for consuming a producer service, expectations and the terms and conditions of use, etc. and the format(s) for the request and response supported, required and offered by the producer.
So far this is simple, there are plenty of web sites that provide static and/or dynamic content, aggregator web sites, web services that provide content / information and mashups that combine and consume the products of multiple producers to provide a new whole...then the complexity can start in, if one is so disposed to allow it.
Sometimes the interactions between consumers and producers have formal definitions, possibly one or more alternatives. For example, a producer may provide a URI to which one may simply send an HTTP GET, or have a description that describes how to format an HTTP POST request or maybe it is an RSS or ATOM feed at this URI, maybe its more formalized....you may have a set of parameters and values, or an XML document described by XSD and/or where you need to authenticate via HTTPS or maybe its a SOAP endpoint whose address is defined using a WS-Addressing endpoint reference, whose operations and requirements are described in WSDL, according to WS-RF or WS-Transfer or both via WS-Resource Transfer (are you with me on the growing complexity?!) with a formalized SLA defined by some other standards, with transactions defined by other standards, with parameters and formats defined by several schemas and namespaces and resources and .... and.... and...Whew!
Now, I'm all for consistency and clearly defined descriptions, definitions, formats, etc. but at some point all this complexity adds up: adds cost, reduces performance, increases error potential, becomes more difficult to use, etc. all of which overshadows the whole reason why a consumer may have wanted to consume the producer's product in the first place....to get something of value done. That value may be profit, learning, etc., but it is useful work. If I have to put a dollar in and my ROI is twenty cents, why would I do that? That's like making charitable contributions to the Red Cross where much of what you donate goes to pay top executives, not to the people who actually need it, you know the one's you're making the donation to help!
So, why work so hard and get back less value than you put in? If you go down the standards-based SOA web service path...well, there you are and there goes your ROI!
If you want more consumers to consume your produced products, you need to provide access as easily and cheaply as possible... so that it is consumable by more consumers. it goes without saying that your information product must be accurate, correct, precise and valid! If your producer is not easy to use then someone else will provide a product that is better, cheaper, faster and/or easier and get used by more consumers....if a consumer needs a CS PhD to use your producer, to consume your content, well there are fewer CS PhD's out there than entrepreneurs, high school students, Moms, Dads, Little League teams, and small businesses and other types of information consumers...you need to make it easier for anyone/everyone...so 3 simple rules
Rule #1: Simple is Better than Comlex. Always.
Rule #2: Simple does not imply less rigor, just less work i.e., work smarter not harder.
Rule #3: If its getting more difficult or complex you are probably doing something wrong (or you have a CS degree and just can't help yourself. I have a CS degree and I often need to remind myself of these rules!)
Perspectives: Simplicity vs. Complexity
Survey the Web. Look at all the most successful information producer / consumer / mashups out there and see what they are using, what their technologies are, where the information they consume comes from and the form of the produced/consumed information. Now, look at what enterprise IT shops are doing, just do a search on jobs for “web services, SOA, architect, enterprise” or look at IT job posting for the Fortune 1000 companies...better yet go read some specs at oasis now put these specs to good use in a simple, easy to use (produce/consume) solution.
What you will find are the two basic camp mentioned before, that I call: 1) the simplicity camp and 2) the complexity camp.
First, I'll describe the complexity camp: Many in the complexity camp work in enterpriise IT. The complexity camp is exemplified by / are big companies with big IT budgets and staff that do much processing over the Web, both internally within their organization and externally with customers, partners and suppliers. They have a lot of legacy systems made up of all those prior technology efforts that linger on, seemingly forever...there is never time to do more than keep them on life support so they in turn continue to provide a source of complexity that grows over time. The complexity camp likes standards. It likes standards so much, that it wants lots of 'em. Closed, old, new, contradictory, overlapping, vague, constrainging and open standards. All these standards then require complex products that are used to create solutions using all these myriad standards. Many of these standards have alternatives from other standards bodies that do some of the same and some different things, so you need people on staff to help you navigate all the different standards, their status and their future direction and to provide coordination of all these standards, etc. Once you have that figured out, you'll need to add your own corporate, division and area standards and interpretations, maybe some SixSigma or other fifth element processes and procedures, and of course more staff to oversee this too. Next, you'll need to have some heavyweight tools to develop solutions using all these standards too. Yes, there are some open source tools and technologies you can use, but to be most efficient you'll probably need to buy some commercial tools too, and these will cost you time and money, maybe small money, maybe big money...it all depends...on what you want to do, and how, and why and how frequently, and other factors. Then of course, you'll need weeks of education to learn how to use those tools. Finally, you are prepared to develop your customer solution. What were those requirements again? So the complexity camp is often found in traditional, enterprise IT staffed with the alchemists of the BS in CS variety where: Complexity is enterprise IT's rationale of existence; just business as usual.
Ask any CEO or business executive if they would be interested in reducing the overhead, costs, size, etc. and/or improve the responsiveness, ROI and effectiveness of IT. Care to guess what their answer may be? And so many people in IT wonder why IT is being outsourced and off shored? Hello! Hello? Anybody home?
Like so many human organizations, that become self-fulfilling and self-sustaining, carrying on the work, form and vestiges of the past long after it ceases to be required... non-profits to cure incurable diseases, civil rights organizations that now advocate for preferential treatment rather than actual equality, etc., Not all, but many IT shops are like this too. Most enterprise CIO's and IT shops are like dinosaurs that haven't realized they are extinct! Case in point, what enterprise IT shop has had to deal with the growth rates and capaciaty expansion experienced by inventors / practitioners of Web 2.0 concepts such as Google, YouTube and MySpace? Few if any! Nor could their IT shops assimliate such growth / speed if they had to! They would collapse! Case Closed.
I make no attempt to hide where my sympathies are, just look at the reality of the Web today, who are the players, the movers, the shakers and the leaders? Web 2.0 companies! The Simplicity Camp is the only realistic, practical and pragmatic way to go if you want to provide viable solutions in a cost effective manner.
The Simplicity Camp
Tim O'Reily initially defined Web 2.0 and many others have adopted and expanded the concept, much is being written about it and more importantly, much is being done with it now, today!
Driven initially by constraints (money, time, other resources, etc.), the simplicity camp looked for alternative ways to approach the problem of web computing, doing more with less. Tried and true technologies of the Web such as HTTP, HTML then XML*, XHTML, etc., low/no cost such as open source software stacks such as LAMP, MAMP, SAMP or even WIMP, then more open source emerged that could assist the simplicity camp, simple more efficient technologies and approaches often based on the tried and true technologies of the Web: HTTP, URI, XML, Namespaces --> and a seminal dissertation by Roy Fielding ---> REST and with the advent of WEB 2.0 a new, more simple way was defined.
The emergence of complementary technologies such as RSS/Atom, Wiki's and AJAX, and producer services available from some big companies in a simple, consumable format i.e., one did not need to have a CS degree and arcane alchemic knowledge of SOA web services, i.e., SOAP, WSDL, UDDI, and more WS-* specs, standards in order to consume these offered services from Google, Amazon, Yahoo and many other well-known industry giants. In fact, more consumers consume information from producers using plain old HTTP GET/POST, REST and/or Web 2.0 than via the standards based, enterprise SOA Web Services. Recently, I heard Google was "sunsetting" = phasing out, its SOAP API for search in favor of an AJAX API (AJAX is a cornerstone Web 2.0 technology which has been so successful that commercial and open source vendors of web services based SOA products have added it to their products! Wonder what that really says?! )
While the standards morass around web services based SOA plods along meandering down its many political paths, the REST / Web 2.0 proponents are providing simple, elegant solutions that enable the web for all to consume, produce and more.
If you're in an entrprise IT department doing SOA and Web Services, and "Enterprise" and/or "Architect" is in your title, you need to have a look around outside your walled garden.
Which camp are you in?
Subscribe to:
Posts (Atom)