 |
|
|
| Re: Some thoughts on VWG... |
|
|
SplitMerge wrote
Hey Everyone,
I don't usually post things on forums, but after spending a few weeks developing an pretty complex application using VWG I feel there are a few things worth chatting about.
|
Very nicely written. I felt like that once too...
I would agree and disagree with some of your points. It's been a while since I've used VWG so if some of my comments are incorrect, by all means, shoot me down and correct me.
>> Stateful objects << I agree this is one of the more major points that helps VWG increase productivity. But in many ways, it's also a major hinderance. Now a website breaks if the server is bounced since the sessions are lost. Scalability is more of an issue. If a machine hosting a properly designed stateless asp.net website can handle a 10000 users, there's no way it will handle the same number of VWG users. A VWG website is simple going to impose more load on the server. I don't think VWG could really be used for popular public websites. But there is no way around this, and that's fine. One just needs to remember that they're getting a huge boost specifically because at the core, they're building a stateful app.
In some ways, I actually think it would be cool to have some concept of a STATELESS VWG app. But this is probably not possible ??
>> Screen Management << The issue here is it's a completely different type of screen now, not one that is really seen on the public web. I've yet to run into a public domain website hosted using VWG. Seems like VWG is best designed for internal business apps that want a WinForm look and feel. Is VWG's own public website usign VWG? Not any other forum, blog, news, whatever that I've seen is VWG. And yet their monthly newsletter brags of 35000+ apps. More apps then registered developers ??? And yet the forum only gets a few dozen posts at most any given day all these years?
>> Messaging needs to focus on Enterprise issues << Amen. Well said.
>> The tool need to project stability, reliability and viability. << More than one client expressed worry about what happens in 5 years. The reply for many years now is basically always "stay tuned".
>> Third Party partnering << Without a doubt. This is one of the major reasons I finally had to abandon VWG. Clients did not like the VWG controls. Every single one had an issue with one control or another and I was helpless to solve the issue unless I paid money to VWG.
Like Navot, said, it seems like VWG touched you in a good way. Same with me. But ultimately, after years "trying" to get a VWG product out, I had to drop it. Last client finally called it quits a while back. Without a doubt, productivity is incredibly increased. But what's missing is the ability to finish an app, polish it off. Back then for me, there were just too many little issues and it simply took the VWG team too long to get fixes out. And each upgrade was a stress nightmare, fix one thing, seven other things break. With VWG I felt like I'd climbed Mount Everest at the speed of light, only to slip, fall and die in the last ten feet. And in less time, I was able to get a Silverlight app out. But Silverlight is without a doubt, not any easy path either, there are frustration there too, mostly the async nature of doing calls and the events mechanism. But it was much much easier to find acceptable controls or even write custom controls.
I had another project come up a while ago and I revisited VWG, but it seemed like little progress has been made on the core VWG and that Gizmox simply continues to expand out from an unpolished core. Still just 6.4 ??..... Cloud, Cloud, uuggghh. Getting old, VWG always chasing the latest buzzword. That aweful datetime picker that every one of my clients hated. No true VWG 3rd party controls. Writing a native control not a trivial task. Still reading of lots of problems people running into with framework. Still lots of breaking changes posts. I just felt that I couldn't afford to spend more time getting back into it to only be let down again. For some people, VWG represents a major risk. Since this project was public facing, I just couldn't take the risk.
However, asp.net, asp.net mvc, jquery, etc, have all come a long a way too and building a good asp.net app today is nowhere as hard as it was 5 years ago when I first tried VWG, all the while, one then still has greater control over their app. When I run into an issue with silverlight, asp.net, jquery, etc, almost always a solution is just a quick google search away. With VWG, it was spend time building a test case to submit. Or digging through their source code only to find a stubbed out function. There is no help for VWG except this forum.
Don't get me wrong, I still have incredible hopes for VWG. One reason why this is still the ONLY product I don't use that I still keep tabs on. I really really really wanted to be able to use it again. I have no doubt I would have made very quick progress. But I just can't afford to fall and die again near the end.
No doubt the reply will be, trust us, the framework is much better now. It's a huge leap of faith...... unfortunately, I'm the one at risk of losing.
|
|
|
|
 |  |
|
|
| Re: Some thoughts on VWG... |
|
|
Hi Happyfirst and thanks allot (once again :)) for the feedback, as you know, we are very open to constructive thoughts!
A few comments/observations on the issues you have raises:
- Scalability – it is a proven fact that VWG is dynamically and unlimited scalable. VWG does handle scalability by providing:
- Minimizing the state into less than 1/10 the state size of a windows app.
- Much improving serialization algorithms with innovative solutions.
- Flattening the graph of objects into an array to further boost the state’s serialization.
Given the above optimizations, VWG can handle hundreds of concurrent users per single server and can scale up to as many concurrent servers on a web farm with less than 15ms “penalty” per call.
This way VWG was selected by MS as an exclusive Building Blocks partner with the MS Windows Azure team. Without unlimited and efficient, dynamic scalability, this couldn’t have happened.
- Public websites – well, you are partially correct, VWG is not the best choice to select as a website’s technology… its targeting is line of business and data-centric applications rather than informative websites or minimal interactive websites like e-commerce and forums.
Anyway, this is not as a result of state-full or not being able to handle tens of thousands of users… VWG does handle today very large apps with up to hundreds of thousands of users, and on clouds!
- Web 2.0 UIs, to save words; you can simply visit http://companionkit.visualwebgui.com for a complete Web 2.0 experience, including scrollable pages and so forth. VWG is heading towards providing rapid tools to create this type of Web 2.0, with the same ease of developing VWG Win-like apps. BTW, the theme designer is only one step which enables “dressing” the app differently. But there’s much more to come…
- Stateless vs. State-full… well, here again, there’s no clear cut… stateless has its limitations and state-full has its limitations of course. Anyway, by dramatically minimizing the state and storing it in any type of intermediate media like: DB, MS State Server, MS Azure App Fabric Cache and any type of rapid storage in clouds and grid hosting minimizes the cons of being state-full to the bare minimum, while enables all its benefits.
- Third party controls (or in its wider name echo system) – By the end of 2010 VWG’s engineers were all heads down to complement missing capabilities and stabilize the framework into its first release version and I believe that now, it’s very stable and so does feedbacks from the community indicate. Gizmox has started to boost the efforts to actively encourage 3rd party vendors to create VWG controls suits. I must say that this effort got allot of traction which will show first signs very soon.
In addition, we believe that the last move towards enhancing the client engine and basing it on jQuery, much improves the community’s ability to extend VWG and create great exciting controls with the much widespread jQurey knowledge that web developers love to use.
- FYI: Gizmox website currently contains a DB of > 60,000 validated, registered developers (both paying and open-source users), some represent teams of a few developers and about 1/3 have developed more than one app already. The number 35,000 applications is only the amount of applications deployed in production by 2009 (latest test was back at early 2009 – I guess the real number today is > 50,000). Furthermore, as far as we know, there are at least twice as much apps being developed as we chat...
- Last and indeed least :) - Gizmox is a relatively small size provider with an innovative approach playing in the field of the giants, the only way for this unique framework to distribute as it does, is by providing solutions for the cutting edge, latest and greatest technologies.
BUT – don’t worry too much, good news about Gizmox/VWG will be announced sometime soon.
Hope this helps!
Kind Regards,
Itzik Spitzen |
|
|
|
|  |