>Hi Fabio,
>
>>>Maybe a language thing, but I don't understand what practical problem you have with SCX. Can you give me a practical situation where the wrong implementation shows ?
>
>Hmmm, that is an impressive list.
>
>>- de have this thread lock bug
>
>Can you explain?
>
In this thread, Geoff reported an error I encountered some months ago.
If You create a datasession class and use it in a form (with DeClass), after the form's execution the datasession class remains instanced and it's impossible to re-open it in "Class Designer"
>>- de have a bug, which casually exchanges the code among the methods
>
>I must admit that I don't use the DE a lot. Most of the time I open the tables via a function in the LOAD or BeforeOpenTables events. I've not seen this problem, at least not in VFP8 SP1.
Other UT memebers reported this error.
>
>>- de have a hard issue:
>> if you use AutoOpenTable,
>>when Opentable/beforeOpentable are executed, the scx properties they have not been loaded yet
>
>Which is AFAIK, by design. I normally stay away from programming those events, but I wonder why you want to address the scx properties in the beforeOpenTable ?
Walter, don't worry why, this design is wrong.
When someone designs a production ambient, he can't assume something isn't worth to be created; because he doesn't often see any interesting application project schemes.
.NET has the same design problem as well.
This could be passable, but only if the desisgner is authoritative and enough brave to change matters in future versions, although this could an incompatible ambient to previous versions.
>
>>- de have another hard issue:
>> with declass, the de becomes readonly,
>>then you cannot build a de class with 3 cursoradapters ( like 3 lookup tables )
>>and add the specific form cursoradapters within the scx.
>
>Looks like a bug to me. To be honest I stayed away from the cursoradaptor.
>
>>- de have others bugs
>
>>- scx.resettodefault() doesn't restore the scx property
>
>I always wondered why anyone wants to use this? I did not find a purpose for it.
>
>>- if you use grid/pageframe/... memberclasses the members.Init isn't executed
>
>Funny. I only use a memberclass for columns in a grid. And luckely they do not contain an INIT event.
>
>>- if you use grid/pageframe/... memberclasses on the class members.Init the name is not
>>that assigned into the scx form.
>
>Not sure what you mean by that.
>
try and put:
MESSAGEBOX(THIS.NAME)
>>- scx have a different events sequence in some situations
>>- vcx classes have various bugs, scx add others
>>- if i design a form, and i want design a variant i have to duplicate it
>
>Well that is why classes do exist, isn't it ?
>
I don't agree.
A scx must only be an instance of the designed class: all the events and their execution order must correspond.
>>- is impossible replicate a scx into a prg.
>
>You mean the restrictions in the programming language for nested containers? Well this applies to visual classes as well. I always found it a bit of a pitty there is no convenient way to go from PRG to VCX, SCX and the other way arround.
>
The difference is that is possible to convert VCX from PRG (they're not the same) and vice versa, while there's no way to define a SCX into a PRG.
>>certainly, writing by chance code case is possible to use the scxes,
>>but I am working to a project in which the whole application
>>it has a homogeneous operation (I have also eliminated the optiongroups, horrible)
>>because you/he/she is produced by a program
>
>I'm not entirely sure what you mean by 'homogenous operation', but I do see that depending on the design strategy you choose you could run into these kind of problems. However, that beeing said, there are alway multiple ways of achieving your goal. In my case where I have been programming for almost 4 years now on one product, I achieved a huge degree of data drivenness that eases the pain of a lot of these kind of issues.
>
>I try to avoid SCX as well, but more from a diffent point of view. I'd like to have most forms kept together in classlibraries. Also, I have a few SCX forms that can be dressed up by a variaty of classes to provide different form functionality.
>
>You show to have a very deep understanding and knowledge of VFPs internal workings. I think you should reconsider the offer of Microsoft to become a beta tester. The fox community needs some guys like you. Esspecially if I read the statement of Randy Brown, I think you could play a role in the future of VFP. I think the community in large would appreciate it. Don't let you be dissapointed by people who have a narrow minded view on what you have found or believe.
>
>
>Keep up the good work.
>
>Walter,
thanks,
Merry Chirstmas and happy New Year
Fabio
