Leonardo Quijano Vincenzi wrote:
Dennis Fleurbaaij wrote:
Leo,

Agree on the Hibernate part, although it's kindof the de-facto standard for database access I guess. And most people don't use Tacos just-for-fun but will need to implement it in larger environments so this might be a good idea. On the other hand, if you strip the annotations it's just a POJO, so it's kindof irrelevant.
I use Hibernate everywhere - that's not the problem. Is rather a problem of minimizing dependencies for Tacos demo. It should be as light as possible. Of course we can always use HSQLDB to avoid asking the user to create a database... but if Pojos work fine, what's the point?
I totally agree, but if we're going to make this just leave it in as comments? So that people can easily relate this to their own models and get it working faster. If we just fill out some of the objects as a trees with parent-child relations (like what you would get back from your EJB layer) it'll be more then enough.
The problem is that I cannot get it work (obiously ;)), debugging output is very minimal at best and when things blow I'm having a very hard time finding things out. This is not really to bad if you have a simple example or very elaborate documentation, but I'm getting the feeling that I'm thrown in the deep end and my swimming skills are somewhat limited. As the standard example uses fileaccess, I can use that a bit but it's not simple enough for me I guess.

So I guess what's lacking is nothing, it's too complete. I can see two main usages for trees, that is filesystems which are are allready in the demo, and data from databases, for which there is no plug-and-play example. I guess a simpler example which addresses the last part would surely help the acceptance/usage of Tacos (atleast overhere).
I agree with you that the demo needs some work, clarifications, etc. It's faster with code contributions, but anyway, it's on my to-do list ;)
If your todo is the size of mine I'll be happy to see it around june ;)

As a matter of fact, I vote for an even simple approach - no file system or database access - for the example. If the examples and docs are clear, you should be able to wire it up in an Hibernate application ;)
No I don't really agree here. If we have the 2 main-use examples people will figure it out. I'm guessing here that most of the tapestry developers are males, we -by nature- try to tinker with it first, destroy it in the process, and _then_ read the docs :) I think that the 2 examples, abeit with inline docs, will be good enough. People that have very specific demands next to this will have to figure things out anyways, but if we provide a simple model to fit their problem to (ie. the POJO's) it'll be simple enough.

Kind regards,
Dennis Fleurbaaij

Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Reply via email to