Comments on Tapestry Central: Announcing Tapestry 5.2

I am currently finishing to read the Alexander Kol...

2011-01-04T07:47:04.819-08:00

I am currently finishing to read the Alexander Kolesnikov's book "Tapestry 5". I have yet to read the final chapter 8 (about creating custom components) to finish it.
It leads me to caracterize Tapestry according the following idea:
(1) DDD: Tapestry does not separate Form and Action into different classes like Struts
(2) page-centric
(3) convention over configuration: for binding purposes
(4) code what you need: one needs to code, for example, an onActive() method in a POJO if and only if one needs to catch the 'activate' event.
Well, Tapestry looks like definitely more high-level than Struts.

Point (4) is about Tapestry easy-to-use testimonial, and puts me to think Tapestry much more as a DSL than a framework (good point).

But I have been disappointed to learn an event handler could not return an instance of a POJO to go to the next page. As a consequence (!), it looks like unfortunately that data are transmitted between different POJOs through a persistence mechanism on top of the http session (as far as I have read correctly the book).
(a) after having reading this book, I just remember the http session looks like the main way to pass data between POJO: it looks like there is no use of the lightweight setAttribute() method on HttpServletRequest object. So, it looks like Tapestry has much more impact on the http session, and challenge more the use of this session, than, let's say, Struts.

And, while the above book has been the single (?) english book for Tapestry during a long time, and while you have been a reviewer, I just want to point some facts that have impacts about Tapestry's evaluation:
(b) there is a single (!) mention of generated JavaScript (for the client-side), page 168
It appears somewhat some (?) controls are generated in JavaScript, no more.
While JavaScript generation has impact on client/server roundtrips and then, on the number of requests sent to servers, it's unfortunate such a book does not dig into this subject.
(c) all examples are about (up to chapter 7, I don't know about the final chapter 8) total page refreshs.
There is no example of partial page refresh.
(d) I would have appreciate a master/detail example, with details displayed into a layer.
Indeed, while Tapestry appears as page-centric, I expected to see an example involving layers to see how well it fits (while I have little doubts Tapestry could handle layers).

Why have I written few points above ?
In fact, I just want to tell you I see Tapestry as a more high-level web framework than Struts, and then, I see it helps to be more productive.
However, as I build web sites targeting a high-level workload, I still have to give answers to the points above (to have a more detailled opinion on Tapestry, and to be able to tell my managers I could rely on it) ; the points are less high-level, but still important for high-performance point of view (Devil is into details), and I have the feeling that, while Struts is too low-level and Tapestry much more high-level, there is a gap between them, and I have not a clear view on how Tapestry fill this gap.

So, I see unfortunate the main book available these last years (250 page) does not give these infos, while this book is quite *very* well-written on the other subjects!
And then, while I think Tapestry could give an (for me, yet-to-know) appropriate answer to these points, I still think more documentation would be appropriate to let me more appreciate your baby and to let it walk much more better for other people's projects. It desserves more complete books (docs?). I hope "Tapestry 5 in Action" would be help here.

My 2 cents.
With regards.

that all sounds very well, especial the documentat...

2010-12-27T07:01:11.227-08:00

that all sounds very well, especial the documentation! In the past it was some time very expensive to find how some thinks can be implement :)

Would you add @PageReset User Guide > Page Life...

2010-12-23T10:48:46.447-08:00

Would you add @PageReset User Guide > Page Life Cycle ?
Some colleagues had hard times finding this annotation.

Note that I was talking about something like putti...

2010-12-22T05:35:17.394-08:00

Note that I was talking about something like putting web site's docs into a global PDF, not looking for "outside" doc ;-)
This being said, I am not against buying computer science books, I have already bought few of them this year...

While "Tapestry 5 in action" is going to be released, is Alexander Kolesnikov's "Tapestry 5" going to be outdated ?
Thanks

This will be available fairly soon as "Tapest...

2010-12-21T09:39:14.757-08:00

This will be available fairly soon as "Tapestry 5 in Action".

Is it possible to gather all (existing) Tapestry d...

2010-12-21T09:23:45.362-08:00

Is it possible to gather all (existing) Tapestry documentation pages (if they make sense as related) into a big single PDF (with, if possible, a smooth learning curve) ?
Thanks.

Well, I like web pages, but when learning a new technology, hypertexts could bring trouble while trying to follow a logical learning line. Let's give you another example: when I read a big PDF doc, if some info is missing, producing a "gap" into the learning curve, then I know there was a pb during writing, and I am then able to submit on a ML my opinion about it. On the other way, when I read web pages, if something is missing, I first wonder I have missed some info/hypertext into the web site, and I start digging into the web site, looking for it: first, I may loose time, and secondly, I may get frustated because I may not be able to know if I missed something myself, or if the info is really missing into the web site.

Well, the other point is that I find too more convenient to download once a big PDF, and reading it printed, for example, during transportation, than printing different web pages and putting them together...