Tuesday, April 03, 2007
Future of Content Management
Critiquing of a recent posting by John Newton...
John posted the Future of Content Management where he posted interesting phrases worthy of deeper analysis:
The Alfresco executive team and I have been on the east coast at a user advisory panel for our American customers here in New York.
This begs the question as to whether advisory panels should solely contain existing paid customers or would there be lift if you had also invited customers of competing products to get their perspective on what it would take to switch.
They also said that none of the other ECM products are thinking about what comes next or at least they aren’t communicating it to their customers - probably a combination of both.
Sadly, I have heard the same conversation from Enterprise Architect peers in other shops that their ECM vendors do a fugly job of communicating their product roadmaps. Many peers have asked for a roadmap that outlines how ECM should ideally integrate with BPM and Enterprise Security. Being that Alfresco is an open source company, maybe you could publish your own roadmap that demonstrates these to convergence points in your next blog entry?
but also pointed out the continuing need to demonstrating trust through workflow for regulated content.
This feels like your customers want to see capabilities of integrating with BPM engines in a standards-based way. Look forward to you describing the best way to integrate with players such as Pega, Intalio, Lombardi, etc
I talked about standards as well. I have been reluctant to blog about these, because I am not sure what the confidentiality rules are around participation in the JCP (the JSR process) or in AIIM’s iECM
I am confused as to why these standards matter. In fact, I believe they propagate a bad habit that needs to stop. In the ECM world, why should every single enterprise reinvent internally developed interfaces when SOA is so pervasive? How difficult would it be to define standardized WSDL for ECM? The model shouldn't be too hard, after all, one can search, view, retrieve, etc. Not much to ECM so why can't it simply come in the box instead of me writing Java code all the freakin time?
The only pushback I got was that one customer believes that cores are low and that we will see 64 cores in this timeframe.
I have several thoughts on this. First, have you ever heard of Azul Systems which makes an appliance that has 384 cores? Your customer is an idiot for arguing about core count when they should be asking of the underlying architecture of your product could even utilize so many cores in a production setting. Do you know that most J2EE application servers can't really scale past eight cores? Do you know there are spots in many open source ECM where techniques such as being syncronized remove the ability to utilize multiple cores? Have you checked your own codebase for scalability characteristics (distinct from performance)?
Someone mentioned Google per content model, but most seemed to agree that this was hard to measure and enforce. Expect more discussion in this area in the future.
I get the feeling that your customers are all idiots. Don't they get the point that the notion of trust should be injected into the process? How about a model of paying not for utilization but the value it brings to the business? I bet most folks can't really calculate an ROI on ECM but simply move forward because they simply understand the problem and there are solutions that bring automation to play
| | View blog reactionsJohn posted the Future of Content Management where he posted interesting phrases worthy of deeper analysis:
This begs the question as to whether advisory panels should solely contain existing paid customers or would there be lift if you had also invited customers of competing products to get their perspective on what it would take to switch.
Sadly, I have heard the same conversation from Enterprise Architect peers in other shops that their ECM vendors do a fugly job of communicating their product roadmaps. Many peers have asked for a roadmap that outlines how ECM should ideally integrate with BPM and Enterprise Security. Being that Alfresco is an open source company, maybe you could publish your own roadmap that demonstrates these to convergence points in your next blog entry?
This feels like your customers want to see capabilities of integrating with BPM engines in a standards-based way. Look forward to you describing the best way to integrate with players such as Pega, Intalio, Lombardi, etc
I am confused as to why these standards matter. In fact, I believe they propagate a bad habit that needs to stop. In the ECM world, why should every single enterprise reinvent internally developed interfaces when SOA is so pervasive? How difficult would it be to define standardized WSDL for ECM? The model shouldn't be too hard, after all, one can search, view, retrieve, etc. Not much to ECM so why can't it simply come in the box instead of me writing Java code all the freakin time?
I have several thoughts on this. First, have you ever heard of Azul Systems which makes an appliance that has 384 cores? Your customer is an idiot for arguing about core count when they should be asking of the underlying architecture of your product could even utilize so many cores in a production setting. Do you know that most J2EE application servers can't really scale past eight cores? Do you know there are spots in many open source ECM where techniques such as being syncronized remove the ability to utilize multiple cores? Have you checked your own codebase for scalability characteristics (distinct from performance)?
I get the feeling that your customers are all idiots. Don't they get the point that the notion of trust should be injected into the process? How about a model of paying not for utilization but the value it brings to the business? I bet most folks can't really calculate an ROI on ECM but simply move forward because they simply understand the problem and there are solutions that bring automation to play
