Hi Frans,

We're going to start quite soon (2-3 months away) on a brand new project - using Visual Studio 2005 Beta 2 to start off with and then obviously using the proper version as and when it ships (whenever that will be!) - the project will not be complete for at leats 7 months.

The plan is to use LLBLGen (of course ), but what would be great is if there was a beta version 2 of LLBLGen that we could use in conjuction with VS2005 beta in the early stages - so that when LLBLGen2 ships we wouldn't need to port it all over, and would be able to take advantage of new features in the beta...... Do you have any plans for an early beta of LLBLGen2?

Matt

MattWoberts wrote:

Hi Frans,

We're going to start quite soon (2-3 months away) on a brand new project - using Visual Studio 2005 Beta 2 to start off with and then obviously using the proper version as and when it ships (whenever that will be!) - the project will not be complete for at leats 7 months.

word of warning; myself and my team have been using the VS2005 beta2 since it was release (for both ASP.NET and WinForms) and although the projects we were working only were very small projects; VS2005 was not very stable and crashed frequently. I would advice you to investigate this matter thoroughly as our experience with the product was not a very happy one

I think anyone using beta software for full development is crazy, unless you're writing software to support multiple toolsets (like LLBLGen Pro).

You have to ask yourself 'how much do I really need this new feature?'.

omar wrote:

MattWoberts wrote:

Hi Frans,

We're going to start quite soon (2-3 months away) on a brand new project - using Visual Studio 2005 Beta 2 to start off with and then obviously using the proper version as and when it ships (whenever that will be!) - the project will not be complete for at leats 7 months.

word of warning; myself and my team have been using the VS2005 beta2 since it was release (for both ASP.NET and WinForms) and although the projects we were working only were very small projects; VS2005 was not very stable and crashed frequently. I would advice you to investigate this matter thoroughly as our experience with the product was not a very happy one

That's interesting. I have heard many people, from folks on the .Net Rocks podcast to many blogs that they are running systems (not small ones) on the Beta 2 go live license with nary a problem.

I guess it depends on who you talk to.

BOb

About vs.net: I haven't used it for longer periods than 2 hours on a day, but every time I used it it stopped accepting new characters in the editor at a given time. I then had to close hte ide, restart the ide, load the project etc. very annoying. Some parts are also very slow IMHO, especially the class designer, unusable.

I can't understand how people can write solid software with this beta2, as there are so many things which fall apart. Ok, your basic test app, that works, but a full application ? I doubt you don't run into serious problems. And what's worse: they ARE going to change some api's in the final.

About LLBLGen Pro v2:

MattWoberts wrote:

The plan is to use LLBLGen (of course ), but what would be great is if there was a beta version 2 of LLBLGen that we could use in conjuction with VS2005 beta in the early stages - so that when LLBLGen2 ships we wouldn't need to port it all over, and would be able to take advantage of new features in the beta...... Do you have any plans for an early beta of LLBLGen2?

Betas of a new version require dedicated beta-testers. So people who will beta-test the v2.0 betas when they come out will be hand-picked, at least that's the plan for now. Perhaps we change it and make it a public beta for all customers, as a teaser to upgrade .

Nevertheless, v2.0 will have new features which will take a while to develop, and I want them in, so it is very hard to say when it is done. I rather polish it for a month longer than rush it out the door.

That said, I personally think they'll never make November 7th. That is: they will have a shiny presentation and the marketing department of MS will be there with their shiny faces and boosted ego's but the rest of the world can only feel sad and sorry at that day: very serious issues are postponed on a daily basis, because fixing them will take too much time and can cause new bugs to appear.

There's now a petition for a beta3. I find that a stupid idea, as forcing them to deliver a beta3 is only cost MORE time, they already are on a 'fix everything and get it done'-schedule. Though I agree with the idea behind the petition that we rather want the bugs to get fixed than postponed just because Marketing already booked the ballroom in Vegas in November.

pilotboba wrote:

That's interesting. I have heard many people, from folks on the .Net Rocks podcast to many blogs that they are running systems (not small ones) on the Beta 2 go live license with nary a problem.

I guess it depends on who you talk to.

Otis wrote:

That said, I personally think they'll never make November 7th. That is: they will have a shiny presentation and the marketing department of MS will be there with their shiny faces and boosted ego's but the rest of the world can only feel sad and sorry at that day: very serious issues are postponed on a daily basis, because fixing them will take too much time and can cause new bugs to appear.

I was in France in JULY for a Ms TTT (Train The Trainer) week on whidby and youkon. What MS are doing to meet the November deadline is drop some features (some really nice time savers specially in the winforms designer ). Their reasoning that there isn't enough resources to fully implement and test these features. And even with the carefully orchastrated demos at the training sessions, it was NOT un-common for thinngs to go crazy especially if you don't do steps in a particular order Frankly, I was very excited to start a real-life VS2005 project specially after reading the same press that alot of people are using it with go-live licenses. Now, I know that is not true and postponed all my plans to a more solid release (a RTM at the earliest)

It's not uncommon to drop features to meet a deadline, I mean, you have to ship at a given point, and we do that too. Though what I think is bad in this case of VS.NET is that resources have been put into features which are not that necessary (I call them the 'demo-features', no-one uses an SqlDataSource on a webform, at least, not if you want to make a career in software engineering, though it demos great. Or the class designer test bench. Who will ever use that clunky visual tester, which is so unproductive, it almost feels like the C# add property 'wizard' (must be Ron Weasley) in vs.net 2003! ), though now they're out of time, they cut serious features as well, which clearly shows that the goals were once again to please marketing, not serious developers. Marketing can't work with "we now have operator restrictions on generic definitions!!". They can work with pretty pictures and drag-n-drop coding style. I'm not saying RAD is bad, it can work very well, but if using a designer takes 10 minutes to replace typing one line, it's not that 'rad'.

I've to see how many vendors for controls will launch their .NET 2.0 stuff right on november 7th, especially the VS.NET integrated tools.

Our project just falls short of the VS2005 date. Its probably going to be in early October we start the project in full swing. Given the size of the project, and the length of time we will spend maintaining and adding to the application (years!), we'd be utterly crazy not to go with VS2005!

What it means for us is there may be an initial few weeks/months when we have to use the beta2, but I can live with that. I can even live with the crashes because I know we're starting something that by the time it gets into full swing, will be chugging away hapily on a final Version 1 of VS2005.

Soooo, the good thing about VS2005 beta2 is that we can plan/prototype and start to code the new application, taking full advantage of it - which is why I was asking about LLBLGen 2.0. We need to start designing the "architecture" for the app, and important considerations are obviosly the DL, passing entity objects back up to the PL, remoting, and other typical considerations. Thats where a working knowledge of LLBLGen 2.0 would be really useful. If I knew more about the planned featues/changes of LLBLGen2.0, then this may well affect the key decisions we make now about structuring the app. I'm wary of starting a brand new project to discover early 2006 that feature X of LLBLGen makes doing Y much easier and faster, and then having to go and refactor loads of code!! Maybe some sort of document explaing what the key changes are likely to be and examples of their usage?

Final thought - I still think you should rename LLBLGen 2.0 to something more cacthy

Otis wrote:

Who will ever use that clunky visual tester, which is so unproductive, it almost feels like the C# add property 'wizard' (must be Ron Weasley) in vs.net 2003! ), though now they're out of time, they cut serious features as well, which clearly shows that the goals were once again to please marketing

If you ever go to a ".NET Roadshow" - they love the glitzy visual tools - "Look how easy it is to add a property to a class" - gets a "oooh" from the crowd every time

MattWoberts wrote:

Soooo, the good thing about VS2005 beta2 is that we can plan/prototype and start to code the new application, taking full advantage of it - which is why I was asking about LLBLGen 2.0. We need to start designing the "architecture" for the app, and important considerations are obviosly the DL, passing entity objects back up to the PL, remoting, and other typical considerations. Thats where a working knowledge of LLBLGen 2.0 would be really useful. If I knew more about the planned featues/changes of LLBLGen2.0, then this may well affect the key decisions we make now about structuring the app. I'm wary of starting a brand new project to discover early 2006 that feature X of LLBLGen makes doing Y much easier and faster, and then having to go and refactor loads of code!! Maybe some sort of document explaing what the key changes are likely to be and examples of their usage?

Well, as .NET 2.0 has more code to do things nicely, a lot of code can be removed from the current base classes, making them fairly small, or even unnecessary. One of the main things is that selfservicing and adapter will share the same core: adapter core and selfservicing is a wrapper around adapter. This will make the code easier to maintain. I will strive to have the code be usable with current code, though there are changes which will cause some migration, like the use of generics and partial classes and perhaps the databinding stuff.

I've made some designs already but not much code has been written yet, though that's not that important for now, the 1.0.2005.1 codebase has to be finished first so it can be used as a base for 2.0's codebases for 1.x and 2.0 .NET.

It's key to have 2.0 done a.s.a.p. but at the same time it has to provide our known quality so it will not be rushed. I know every VS.NET 2005 beta2 user wants to get their hands on a beta a.s.a.p. but all I can say it is coming but I won't set a date now nor a fixed featureset.

Final thought - I still think you should rename LLBLGen 2.0 to something more cacthy

Type in 'LLBLGen' at google and you know why that would be a very silly move

Otis wrote:

Type in 'LLBLGen' at google and you know why that would be a very silly move

Thats a very good point Thanks for the info Frans!