WCF
-
Is it just me, or does does WCF seem clunky, overly complicated, and clumsy?
"Why don't you tie a kerosene-soaked rag around your ankles so the ants won't climb up and eat your candy ass..." - Dale Earnhardt, 1997
-----
"...the staggering layers of obscenity in your statement make it a work of art on so many levels." - Jason Jystad, 10/26/2001I thought it was pretty easy and straight forward to be honest, especially when it took me all of 5 minutes to host a WCF service within a windows service.
-
Rama Krishna Vavilala wrote:
The thing I really like about it is that it is pretty simple to remote your classes in different ways just by changing the configuration files.
Is this actually a common need? Do a significant number of people actually develop using one transport/protocol and then deploy using another? What's the scenario there? I mean, when WCF was first announced, that feature was demo'd pretty heavily - "and now, by changing two lines in the app.config, we've switched from SOAP to REST for our protocol, and HTTP to carrier pigeons for our transport - and the client still displays 'hello world'!" It was cool. I clapped. Then i went home, back to cursing the SOAP layer for not working with the non-MS SOAP server that i need to use. And i just can't help but think that the flexibility was built into the wrong layer.
Citizen 20.1.01
'The question is,' said Humpty Dumpty, 'which is to be master - that's all.'
Shog9 wrote:
Is this actually a common need?
For me yes. I wanted to support windows authentication and a custom authentication both together. No code changes simply expose the end points. For LAN users, I allow TCP and for users outside a firewall I allow WS*. It all worked well. Considering tons of code I wrote in the DCOM version of my application to support custom security and custom marshaling, WCF is certainly a welcome addition. The disclaimer here is that the product has not yet been deployed on actual customer site but in our test environment with or without firewall it worked well.
Shog9 wrote:
with the non-MS SOAP server
Aha.. Interoperability has always been a pain with any WS technology. Incidentally, that is what WS were meant to solve. For J2EE interoperability, there is a project tango. http://java.sun.com/developer/technicalArticles/glassfish/ProjectTango/ Anyway, I will not blame WCF for interoperability issues. We need a CPHog for Safari:)
-
Shog9 wrote:
Is this actually a common need?
For me yes. I wanted to support windows authentication and a custom authentication both together. No code changes simply expose the end points. For LAN users, I allow TCP and for users outside a firewall I allow WS*. It all worked well. Considering tons of code I wrote in the DCOM version of my application to support custom security and custom marshaling, WCF is certainly a welcome addition. The disclaimer here is that the product has not yet been deployed on actual customer site but in our test environment with or without firewall it worked well.
Shog9 wrote:
with the non-MS SOAP server
Aha.. Interoperability has always been a pain with any WS technology. Incidentally, that is what WS were meant to solve. For J2EE interoperability, there is a project tango. http://java.sun.com/developer/technicalArticles/glassfish/ProjectTango/ Anyway, I will not blame WCF for interoperability issues. We need a CPHog for Safari:)
Rama Krishna Vavilala wrote:
We need a CPHog for Safari:)
There's always Firefox for the Mac. ;)
-
Rama Krishna Vavilala wrote:
We need a CPHog for Safari:)
There's always Firefox for the Mac. ;)
Yes but I don't like it as much as Safari on Mac.
You have, what I would term, a very formal turn of phrase not seen in these isles since the old King passed from this world to the next. martin_hughes on VDK
-
Is it just me, or does does WCF seem clunky, overly complicated, and clumsy?
"Why don't you tie a kerosene-soaked rag around your ankles so the ants won't climb up and eat your candy ass..." - Dale Earnhardt, 1997
-----
"...the staggering layers of obscenity in your statement make it a work of art on so many levels." - Jason Jystad, 10/26/2001Nope. I like it - I actually find it a better fit for remoting purposes, please it's relatively easy to secure. The biggest problem is a lack of decent articles on it. I started writing one on using it to communicate with Amazon S3, but packed it in when I was spending more and more time clunking round the Amazon API, rather than demonstrating "cool" ways of doing things in WCF. It may be time to revisit this, especially as Visual Studio 2008 makes creating and consuming WCF apps much easier.
Deja View - the feeling that you've seen this post before.
-
Shog9 wrote:
Is this actually a common need?
For me yes. I wanted to support windows authentication and a custom authentication both together. No code changes simply expose the end points. For LAN users, I allow TCP and for users outside a firewall I allow WS*. It all worked well. Considering tons of code I wrote in the DCOM version of my application to support custom security and custom marshaling, WCF is certainly a welcome addition. The disclaimer here is that the product has not yet been deployed on actual customer site but in our test environment with or without firewall it worked well.
Shog9 wrote:
with the non-MS SOAP server
Aha.. Interoperability has always been a pain with any WS technology. Incidentally, that is what WS were meant to solve. For J2EE interoperability, there is a project tango. http://java.sun.com/developer/technicalArticles/glassfish/ProjectTango/ Anyway, I will not blame WCF for interoperability issues. We need a CPHog for Safari:)
Rama Krishna Vavilala wrote:
For LAN users, I allow TCP and for users outside a firewall I allow WS*. It all worked well.
Ok, that makes sense (assuming you obtain greater efficiency for the common scenario then at least).
Rama Krishna Vavilala wrote:
Interoperability has always been a pain with any WS technology. Incidentally, that is what WS were meant to solve.
Ha! Yeah, the server's a third-party CGI deal, pretty sure all the SOAP stuff is done using some C++ library. It's buggy - there are encoding issues, and some datatypes that don't play well with the MS stuff. The frustrating part of it all is that so much of the stack appears to be implemented as internal or sealed classes, meaning i have to put hacks in at a much lower level than i'd rather deal with.
Rama Krishna Vavilala wrote:
Anyway, I will not blame WCF for interoperability issues.
Eh, it didn't start with WCF. It just didn't get any better from my perspective.
Rama Krishna Vavilala wrote:
We need a CPHog for Safari
You volunteering?;P
Citizen 20.1.01
'The question is,' said Humpty Dumpty, 'which is to be master - that's all.'
-
Nope. I like it - I actually find it a better fit for remoting purposes, please it's relatively easy to secure. The biggest problem is a lack of decent articles on it. I started writing one on using it to communicate with Amazon S3, but packed it in when I was spending more and more time clunking round the Amazon API, rather than demonstrating "cool" ways of doing things in WCF. It may be time to revisit this, especially as Visual Studio 2008 makes creating and consuming WCF apps much easier.
Deja View - the feeling that you've seen this post before.
I find it incredulous I have to manually run external tools to make it go. Plus, I have not background at all in dealing with remoting, so references and comparisons of that technology are wasted on me. You're right, there's nothing DECENT yet written to describe WCF and how to make use of it. It also doesn't appear to be something you can wrap a class around to make it more of a black box.
"Why don't you tie a kerosene-soaked rag around your ankles so the ants won't climb up and eat your candy ass..." - Dale Earnhardt, 1997
-----
"...the staggering layers of obscenity in your statement make it a work of art on so many levels." - Jason Jystad, 10/26/2001 -
Yes but I don't like it as much as Safari on Mac.
You have, what I would term, a very formal turn of phrase not seen in these isles since the old King passed from this world to the next. martin_hughes on VDK
Yeah yeah. You Mac users and your need for all those standardized UI gadgets. ;) To be perfectly honest with you, there's no way that we're going to support Safari in the near-future. The way that CPhog works is that we have a very light UserScript that basically serves to inject the real script into the page directly. This enables us to get away from using a lot of the Greasemonkey-specific functions we'd have to use for things like XmlHttpRequest and the like. However, that's also enabled us to do a lot of cool stuff with globalStorage and Javascript 1.7 and the like that aren't supported in the current versions of any browser other than Firefox. It would be easy to abstract the storage stuff....seeing as how it's in a nice StorageWrapper class I wrote. We've thought about doing that to provide storage persistence for users across machines. But some of the JS, DOM, and CSS features we're using are only present in Firefox. You'd have to really convince one of us that there's a strong Safari use-case here...which would be hard because, for all of the comments about how responsive we are to user requests, we really only care about ourselves. Most of the features in CPhog are there because we get irritated with something and decide that we want to fix stuff. Basically, you're going to have to buy Shog or myself a Mac, convince us that Safari rocks, and get us to use it. Same thing goes for Opera. John Cardinal said that he'd use CPhog if we provided an Opera version of it. I looked at it and decided, in about 10 minutes, that there's no way we're going to invest that time and effort. But you're welcome to branch the code and try to get CPhog working on Safari. I'll even give you svn access if you want. ;)
-
Yes but I don't like it as much as Safari on Mac.
You have, what I would term, a very formal turn of phrase not seen in these isles since the old King passed from this world to the next. martin_hughes on VDK
Have you tried Camino? (Not sure if CPHog works with it though :~)
-- Kein Mitleid Für Die Mehrheit
-
I find it incredulous I have to manually run external tools to make it go. Plus, I have not background at all in dealing with remoting, so references and comparisons of that technology are wasted on me. You're right, there's nothing DECENT yet written to describe WCF and how to make use of it. It also doesn't appear to be something you can wrap a class around to make it more of a black box.
"Why don't you tie a kerosene-soaked rag around your ankles so the ants won't climb up and eat your candy ass..." - Dale Earnhardt, 1997
-----
"...the staggering layers of obscenity in your statement make it a work of art on so many levels." - Jason Jystad, 10/26/2001John Simmons / outlaw programmer wrote:
I find it incredulous I have to manually run external tools to make it go.
I manage to do it all inside Visual Studio now (well, developing and testing that is). VS2008 introduced its own service hoster which is a real timesaver.
John Simmons / outlaw programmer wrote:
It also doesn't appear to be something you can wrap a class around to make it more of a black box.
Well, there are some bits you can "black box", but they aren't hugely useful. What is good is that it pushes you down the route of abstracting the datamodel from the actual service functionality. To that extent, it's a huge step forward.
Deja View - the feeling that you've seen this post before.
-
Shog9 wrote:
Is this actually a common need?
For me yes. I wanted to support windows authentication and a custom authentication both together. No code changes simply expose the end points. For LAN users, I allow TCP and for users outside a firewall I allow WS*. It all worked well. Considering tons of code I wrote in the DCOM version of my application to support custom security and custom marshaling, WCF is certainly a welcome addition. The disclaimer here is that the product has not yet been deployed on actual customer site but in our test environment with or without firewall it worked well.
Shog9 wrote:
with the non-MS SOAP server
Aha.. Interoperability has always been a pain with any WS technology. Incidentally, that is what WS were meant to solve. For J2EE interoperability, there is a project tango. http://java.sun.com/developer/technicalArticles/glassfish/ProjectTango/ Anyway, I will not blame WCF for interoperability issues. We need a CPHog for Safari:)
> wanted to support (snip) and a custom authentication both together > It all worked well Care to share your experiences there? For service-oriented scenarios over the public net, web, etc. that's what I am aiming at as a context of the query... I still fail to see how they address anything apart from supporting CardSpace (okay, they now have pluggable token service support so we can built anything we want but that's beside the point if everyone does it their own way ). Even when LiveID support kicks in (well I hope it doesn't), I don't see a decent way to authenticate anyone without usual, cumbersome and expensive certificate management (it gets worse if you need more than authent, but I'd be happy so see how people approach just that task). (it's all fine and great that it supports multiple auth schemes but I'm still to see one that works/heavily used out there, or more managable than certs)
modified on Tuesday, April 22, 2008 3:17 PM
-
Yeah yeah. You Mac users and your need for all those standardized UI gadgets. ;) To be perfectly honest with you, there's no way that we're going to support Safari in the near-future. The way that CPhog works is that we have a very light UserScript that basically serves to inject the real script into the page directly. This enables us to get away from using a lot of the Greasemonkey-specific functions we'd have to use for things like XmlHttpRequest and the like. However, that's also enabled us to do a lot of cool stuff with globalStorage and Javascript 1.7 and the like that aren't supported in the current versions of any browser other than Firefox. It would be easy to abstract the storage stuff....seeing as how it's in a nice StorageWrapper class I wrote. We've thought about doing that to provide storage persistence for users across machines. But some of the JS, DOM, and CSS features we're using are only present in Firefox. You'd have to really convince one of us that there's a strong Safari use-case here...which would be hard because, for all of the comments about how responsive we are to user requests, we really only care about ourselves. Most of the features in CPhog are there because we get irritated with something and decide that we want to fix stuff. Basically, you're going to have to buy Shog or myself a Mac, convince us that Safari rocks, and get us to use it. Same thing goes for Opera. John Cardinal said that he'd use CPhog if we provided an Opera version of it. I looked at it and decided, in about 10 minutes, that there's no way we're going to invest that time and effort. But you're welcome to branch the code and try to get CPhog working on Safari. I'll even give you svn access if you want. ;)
I guess FF and Safari might be very similar. Though I have not gone into the details yet.
David Stone wrote:
get CPhog working on Safari.
I think the biggest challenge will be to get the user scripts running. I saw a few greasemonkey clones but did not manage to get them working.
You have, what I would term, a very formal turn of phrase not seen in these isles since the old King passed from this world to the next. martin_hughes on VDK
-
Have you tried Camino? (Not sure if CPHog works with it though :~)
-- Kein Mitleid Für Die Mehrheit
Not yet! I searched for a few GM clones but I did not try them as I wanted to build my own GM clone:)
-
> wanted to support (snip) and a custom authentication both together > It all worked well Care to share your experiences there? For service-oriented scenarios over the public net, web, etc. that's what I am aiming at as a context of the query... I still fail to see how they address anything apart from supporting CardSpace (okay, they now have pluggable token service support so we can built anything we want but that's beside the point if everyone does it their own way ). Even when LiveID support kicks in (well I hope it doesn't), I don't see a decent way to authenticate anyone without usual, cumbersome and expensive certificate management (it gets worse if you need more than authent, but I'd be happy so see how people approach just that task). (it's all fine and great that it supports multiple auth schemes but I'm still to see one that works/heavily used out there, or more managable than certs)
modified on Tuesday, April 22, 2008 3:17 PM
Yes, a certificate is needed and I will say that it was the most difficult part. But it was all related to configuration. (I did spend hours researching on how to get the UserName Password work without certificates) As soon as I added a certificate it worked like a charm. You can create two endpoints one with windows auth and one with UserName password auth (with a cert needed). It works!
You have, what I would term, a very formal turn of phrase not seen in these isles since the old King passed from this world to the next. martin_hughes on VDK
-
I guess FF and Safari might be very similar. Though I have not gone into the details yet.
David Stone wrote:
get CPhog working on Safari.
I think the biggest challenge will be to get the user scripts running. I saw a few greasemonkey clones but did not manage to get them working.
You have, what I would term, a very formal turn of phrase not seen in these isles since the old King passed from this world to the next. martin_hughes on VDK
Rama Krishna Vavilala wrote:
I guess FF and Safari might be very similar. Though I have not gone into the details yet.
Not really in terms of their rendering engines. Firefox uses Gecko. Safari uses WebKit, which shares its roots with KHTML. They render HTML and CSS very differently.
Rama Krishna Vavilala wrote:
I saw a few greasemonkey clones but did not manage to get them working.
Yeah. But like I said, there's not a whole lot of Greasemonkey specific stuff we use. Mostly it's the WHATWG globalStorage working draft spec, JS 1.7 stuff, and all the "Hey, it works in Firefox, so it's okay" CSS and HTML we inject into the page. You fix those, you can have CPhog on Safari. ;)
-
Is it just me, or does does WCF seem clunky, overly complicated, and clumsy?
"Why don't you tie a kerosene-soaked rag around your ankles so the ants won't climb up and eat your candy ass..." - Dale Earnhardt, 1997
-----
"...the staggering layers of obscenity in your statement make it a work of art on so many levels." - Jason Jystad, 10/26/2001Just reading up on WCF after your post. Wondering to myself if there is a way for perimeter DMZ web GUI's to talk to web services/databases inside the fire wall. Curious, since WCF seems to rely in part on MSMQ which works only intra-domain and DMZ web servers of course are not part of the inside-the-firewall domain... need to read up some more, thanks for bringing up WCF.
MrPlankton
-
Just reading up on WCF after your post. Wondering to myself if there is a way for perimeter DMZ web GUI's to talk to web services/databases inside the fire wall. Curious, since WCF seems to rely in part on MSMQ which works only intra-domain and DMZ web servers of course are not part of the inside-the-firewall domain... need to read up some more, thanks for bringing up WCF.
MrPlankton
MSMQ is only one of the bindings (a damn handy one granted). Other bindings include peer to peer, pipes, SOAP and so on. Effectively you can think of WCF as following ABC: A = Address - the address of the service you are communicating with B = Binding - the binding for the service (see above) C = Contract - the contract that the service exposes for you to communicate with
Deja View - the feeling that you've seen this post before.
-
MSMQ is only one of the bindings (a damn handy one granted). Other bindings include peer to peer, pipes, SOAP and so on. Effectively you can think of WCF as following ABC: A = Address - the address of the service you are communicating with B = Binding - the binding for the service (see above) C = Contract - the contract that the service exposes for you to communicate with
Deja View - the feeling that you've seen this post before.
Good to know, thanks.
MrPlankton
-
Nope. I like it - I actually find it a better fit for remoting purposes, please it's relatively easy to secure. The biggest problem is a lack of decent articles on it. I started writing one on using it to communicate with Amazon S3, but packed it in when I was spending more and more time clunking round the Amazon API, rather than demonstrating "cool" ways of doing things in WCF. It may be time to revisit this, especially as Visual Studio 2008 makes creating and consuming WCF apps much easier.
Deja View - the feeling that you've seen this post before.
Most difficult thing - deploying the bloody thing (typo in the config file :-O)
Never underestimate the power of human stupidity RAH
-
Most difficult thing - deploying the bloody thing (typo in the config file :-O)
Never underestimate the power of human stupidity RAH
Ah yes. The 50 different references you have to update because you decided that Service1 is a really naff name for a service.
Deja View - the feeling that you've seen this post before.