Shamil Salakhetdinov
shamil at smsconsulting.spb.ru
Sat Sep 11 01:16:34 CDT 2010
Hi Gustav -- Thank you for your extended note and your "first hand" experience with LightSwitch. So, LightSwitch desktop UI is also Silverlight based? (from keynote video I didn't get that) IOW this tool should be better called "Silver-Light-Switch" ? (Does it "smell" a bit as "Silver-Bullet" just based on its title?) <<< Of course, like Access, this is a dangerous tool for those not knowing relational databases - but for you and I and our fellow list members who know about databases, this could be a very strong tool - mastering this could make you very competitive for small projects. >>> Yes, and I'd suppose this tool is also "dangerous" for you and me :) Yes, I see you noted - for "small projects" - but how often "small projects" quickly grow into "big ones" - almost always in my experience, and what about yours? I guess it should be very similar... IMO when working with such a tool one is becoming a "lazy developer". I can be wrong but the "hidden danger" of such a tool IMO is that while one is working using it on a small app it works well but then, suddenly(?!), they hit the wall - and there is often no resources and no will to "break that wall" - and an application degrades steadily, its support costs grow (astronomically), and then it gets abandoned. Everybody here "been there seen that" many times I suppose... Yes, as a prototyping tool is looks good - and prototyping both business datamodel + FE rough interface/navigation/data validation but when prototyping is done - "throw it away", and make "real handwritten software". I can be wrong of course, we will see... But Test Driven Development (TDD) and Domain Driven Design (DDD) is a synonym for professional agile software development for me, and LightSwitch looks so different from TDD/DDD mainstream tendence... LightSwitch still looks more like "VB6 v.2010 on .NET Framework+WCF/SilverLight+EF steroids" with all the dangers of VB6 in creation of "world best spaghetti code"... Yes, the fact is that business software isn't that a rocket science experience with a lot of routine repetitive tasks and code and UI templates "crying for" automation for ages now - still no good results - would we finally get good results now with LightSwitch? Let's hope so, just hope... But the more business software development automation fails to deliver a universal tool the more it's becoming clear that there couldn't be such a tool in principle? LightSwitch is a high level framework on top of several frameworks - such a "vertical structure" of stacked frameworks usually results in a rather rigid construction. Hunting for RAD and flexibility and in fact getting into the opposite direction in long run (for big projects) + becoming a "lazy developer": it was like that before - why it should be different this time? Doesn't "The Mythical Man-Month" by Frederick Phillips Brooks, Jr. (http://en.wikipedia.org/wiki/Fred_Brooks) and the "The Design of Design" of the same author tells us that the real breakthrough in (business) software development automation is yet very far from nowadays if ever happens? The main losses/overheads in (business) software development as it's known are in the area of communicating customers'/stakeholders' requirements to the business analysts/software designers/developers - LightSwitch (as VSLive Keynote speech states) is going to solve/significantly diminish the losses/overheads of this core issue - I doubt it will work well here... Microsoft is investing so much into high level software development frameworks these days as they never did before - is that more "a marketing move" or a long term well thought through strategy - who knows? No doubts Microsoft can afford world best computer science analysts and software developers - did they find how to make a "Silver-Bullet" this time? I can be wrong, feel free to accept all the above as a rumbling... Just found this note "LightSwitch Extensibility: It Ain't Just Hype" http://geekswithblogs.net/andrewbrust/archive/2010/08/21/lightswitch-extensi bility-it-ainrsquot-just-hype.aspx... Thank you. -- Shamil -----Original Message----- From: accessd-bounces at databaseadvisors.com [mailto:accessd-bounces at databaseadvisors.com] On Behalf Of Gustav Brock Sent: Saturday, September 11, 2010 1:00 AM To: accessd at databaseadvisors.com Subject: Re: [AccessD] LightSwitch (was: Automating web page entry (was:Scrollbutton)) Hi Shamil Well, I'm working my way through the tutorial that's for download as well and I'm impressed. Please note the target applications: small business apps with no other purpose than "getting work done". Such apps often have no special requirements for design or features, yet they must be reliable and consistent to use - yet fast to develop or you simply don't have the budget. Most data is collected in a local lsml (LightSwitchML) file but it connects via the EF to nearly everything just like that. It is so flexible, and validation, error messages and so on is ready at hand with zero or extremely little code. Importantly, the EF let's you "remodel" any data source making it very easy to adapt and connect/relate different data sources - again with zero code; this feature alone is worth studying. The default design of screens (forms) is very neat (using Silverlight) and so far from what you can handcraft otherwise (even using hours) with VS. Of course, like Access, this is a dangerous tool for those not knowing relational databases - but for you and I and our fellow list members who know about databases, this could be a very strong tool - mastering this could make you very competitive for small projects. Just today we were assigned a new straight-forward project (controlling projects/employees/contracts/salaries for a small TV production company) where we will put it to use to check it out. /gustav >>> shamil at smsconsulting.spb.ru 10-09-2010 21:33 >>> Hi Gustav -- Thank you for your note. Starting from your link I have got to this August VSLive's KeyNote "A Glimpse of the Future..." dedicated to LightSwitch. And I have watched that almost one hour long video recording. But I'm still not convinced about going to play with this tool. IMO it's now: - a good prototyping tool; - a good MS marketing tool; I will wait for release. Maybe even for version #2 release. Do you know how metadata of this tool are organized? Is there technical documentation describing that metadata? Thank you. -- Shamil <<< snip >>>