Comments for Musings by Seref Arikan http://www.serefarikan.com About healthcare informatics and technology (mostly) Tue, 07 Sep 2010 09:06:45 +0000 http://wordpress.org/?v=2.2.2 Comment on Pulse, holy grail for Eclipse community? by Jens Eckels http://www.serefarikan.com/?p=43#comment-19140 Jens Eckels Tue, 19 Aug 2008 18:50:32 +0000 http://www.serefarikan.com/?p=43#comment-19140 sarikan; Thanks for your post. We're glad you are finding Pulse so useful. Any and all further comments are appreciated, as we never stop improving the product(s). Best, Jens Pulse Team sarikan;

Thanks for your post. We’re glad you are finding Pulse so useful. Any and all further comments are appreciated, as we never stop improving the product(s).

Best,
Jens
Pulse Team

]]> Comment on There is trouble at the front layer! by sarikan http://www.serefarikan.com/?p=42#comment-18195 sarikan Tue, 12 Aug 2008 19:20:57 +0000 http://www.serefarikan.com/?p=42#comment-18195 Hi Brian, Quite interesting work, would you mind giving a little bit background? What is your motivation for your work? I'd like to get your opinions, since you've worked in a different context. HL7 is another beast, one which you can not avoid in many projects. Your statement about what is needed in the client side is hard to answer, but I'd like to see what happens if we assume CUI as what is needed in the client side. Of course CUI does not cover all possible GUI functionality, but still it is a well defined target for connecting to models expressed in domain specific languages like the ones we've been discussing. Hi Brian,
Quite interesting work, would you mind giving a little bit background? What is your motivation for your work? I’d like to get your opinions, since you’ve worked in a different context. HL7 is another beast, one which you can not avoid in many projects.
Your statement about what is needed in the client side is hard to answer, but I’d like to see what happens if we assume CUI as what is needed in the client side. Of course CUI does not cover all possible GUI functionality, but still it is a well defined target for connecting to models expressed in domain specific languages like the ones we’ve been discussing.

]]> Comment on There is trouble at the front layer! by Brian DeCamp http://www.serefarikan.com/?p=42#comment-18012 Brian DeCamp Mon, 11 Aug 2008 22:27:43 +0000 http://www.serefarikan.com/?p=42#comment-18012 I've actually done quite a lot of exploring in this area, although not with CUI. I'm focusing on Flex UIs, and I've ported much of the core HL7 javasig data types and core classes into AS3 for this purpose. I've also created mappings of domain models into AS3 for tighter constraints on the client side. There are some difficulties with this approach, and I have abandoned it for the moment until I have more resources to solve the issues. It is difficult to get a domain model that is constrained enough to be convenient on the client. Many of the datatypes in HL7 have too much logic for client side processing, so it would take some time to get exactly what you want and need on the client side. I think the approach has merit, and it could be used to create some very rich interfaces, but it's a good bit of work. I’ve actually done quite a lot of exploring in this area, although not with CUI. I’m focusing on Flex UIs, and I’ve ported much of the core HL7 javasig data types and core classes into AS3 for this purpose. I’ve also created mappings of domain models into AS3 for tighter constraints on the client side. There are some difficulties with this approach, and I have abandoned it for the moment until I have more resources to solve the issues. It is difficult to get a domain model that is constrained enough to be convenient on the client. Many of the datatypes in HL7 have too much logic for client side processing, so it would take some time to get exactly what you want and need on the client side. I think the approach has merit, and it could be used to create some very rich interfaces, but it’s a good bit of work.

]]> Comment on Can you build a product for data mining? by sarikan http://www.serefarikan.com/?p=31#comment-6246 sarikan Wed, 21 May 2008 07:57:58 +0000 http://www.serefarikan.com/?p=31#comment-6246 Hi, I'd certainly be interested to hear about this. We need to get better at machine learning to make any use of huge piles of data that we've been saving for almost decades now. I really do not want to see all this data as input into another "business intelligence" tool!. Cheers Seref Hi,
I’d certainly be interested to hear about this. We need to get better at machine learning to make any use of huge piles of data that we’ve been saving for almost decades now. I really do not want to see all this data as input into another “business intelligence” tool!.

Cheers
Seref

]]> Comment on Can you build a product for data mining? by Erk Subasi http://www.serefarikan.com/?p=31#comment-6117 Erk Subasi Tue, 20 May 2008 13:47:15 +0000 http://www.serefarikan.com/?p=31#comment-6117 Great points Seref, I know some people out there working on learning bayesian networks on GPUs. I also find this approach quite promising, actually would like to try an implementation personally if I'll be able to find some time. Let you know on any progress. ;) cheers. Great points Seref,

I know some people out there working on learning bayesian networks on GPUs. I also find this approach quite promising, actually would like to try an implementation personally if I’ll be able to find some time. Let you know on any progress. ;)
cheers.

]]> Comment on Microsoft Amalga, why did not anyone notice this? by Seref http://www.serefarikan.com/?p=28#comment-3827 Seref Thu, 01 May 2008 17:47:23 +0000 http://www.serefarikan.com/?p=28#comment-3827 I've been following almost all efforts which can be called DSL based. IMHO OpenEHR is the closest to being called one. The amount of effort that is necessary to build such a framework is enormous. DSL can be considered the focus of a new approach, but it requires a whole set of other layers, like persistence, some form of remoting (web services), multiple GUI technologies... This is a big challenge, and I expect a vendor like Microsoft to tackle it. In fact, I think that there are signs of this approach in their technology portfolio at the moment. I do not know any healthcare vendors with the financial power to build a "complete" DSL based solution. I’ve been following almost all efforts which can be called DSL based. IMHO OpenEHR is the closest to being called one.
The amount of effort that is necessary to build such a framework is enormous. DSL can be considered the focus of a new approach, but it requires a whole set of other layers, like persistence, some form of remoting (web services), multiple GUI technologies…
This is a big challenge, and I expect a vendor like Microsoft to tackle it. In fact, I think that there are signs of this approach in their technology portfolio at the moment. I do not know any healthcare vendors with the financial power to build a “complete” DSL based solution.

]]> Comment on Microsoft Amalga, why did not anyone notice this? by Anonymous http://www.serefarikan.com/?p=28#comment-3105 Anonymous Wed, 23 Apr 2008 08:29:53 +0000 http://www.serefarikan.com/?p=28#comment-3105 If having a DSL would be helpful for these healthcare companies, why do you believe they don't go ahead and build there own custom one to precisely meet there needs, rather than waiting fro Microsoft to provide an 'almost there' solution? If having a DSL would be helpful for these healthcare companies, why do you believe they don’t go ahead and build there own custom one to precisely meet there needs, rather than waiting fro Microsoft to provide an ‘almost there’ solution?

]]> Comment on Netbeans 6 vs 3rd Rail by sarikan http://www.serefarikan.com/?p=11#comment-26 sarikan Fri, 15 Feb 2008 10:31:22 +0000 http://www.serefarikan.com/?p=11#comment-26 My rails work has been in halt for a while, but if I go back, I'll probably use Netbeans for it. I have not come across a better choice since I've written this blog entry My rails work has been in halt for a while, but if I go back, I’ll probably use Netbeans for it. I have not come across a better choice since I’ve written this blog entry

]]> Comment on Netbeans 6 vs 3rd Rail by Thom Parkin http://www.serefarikan.com/?p=11#comment-25 Thom Parkin Thu, 14 Feb 2008 16:31:23 +0000 http://www.serefarikan.com/?p=11#comment-25 I have always been impressed (and pleased) with Borland products. Way, way back to the days of Turbo Pascal and Turbo Assembler! New to Ruby on Rails, I was extremely impressed with NetBeans. With Sun's decision to purchase MySQL I think this might be the better choice for RoR development. The ability to use existing Java (JAR) routines was demonstrated to me and left a lingering positive impression of NetBeans. I am struggling, as you indicated, with 3rd Rail. Although I really want to give Borland a chance, I cannot afford the time it takes to get this beautiful IDE working properly. I have always been impressed (and pleased) with Borland products.
Way, way back to the days of Turbo Pascal and Turbo Assembler!
New to Ruby on Rails, I was extremely impressed with NetBeans. With Sun’s decision to purchase MySQL I think this might be the better choice for RoR development. The ability to use existing Java (JAR) routines was demonstrated to me and left a lingering positive impression of NetBeans.
I am struggling, as you indicated, with 3rd Rail. Although I really want to give Borland a chance, I cannot afford the time it takes to get this beautiful IDE working properly.

]]> Comment on Microsoft CUI, a brave attempt by sarikan http://www.serefarikan.com/?p=8#comment-6 sarikan Tue, 11 Dec 2007 10:18:11 +0000 http://www.serefarikan.com/?p=8#comment-6 yariyor olmasi lazim Tayfun hocam, bu tarafi ayarlamamistim :) yariyor olmasi lazim Tayfun hocam, bu tarafi ayarlamamistim :)

]]>