Those are good observations. Building a system such as Tapestry and improving it for every release is a challenge. In Tapestry's case, the inheritance imposition, and the lack of a clear divider between internal/implementation and public/api has been a growing problem from release to release. Tapestry 5 aims to clear that up and, because of the nature of the contract between the framework and the user code, may realistically clear it up far, far into the future.
I also think that the new design will simplify the learning curve, and I'm committed to building documentation and examples/tutorials as I go. Again, the things that cause minor-to-moderate compatibility problems (between 3 and 4) can only be effectively fixed by introducing a non-backwards compatible change (5). In fact, many of their observations dovetail nicely with my plans for 5. For example, "too flexible" will be simplified by removing .page and .jwc files (and just using annotations). It's nice to know that a challenging, picky (in a good way) client like this is also sold on Tapestry! On 5/16/06, Brian K. Wallace <[EMAIL PROTECTED]> wrote:
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I pondered the right way to relay the following and decided this might be the best. This is not meant to do anything other than relay information, but if you see a correlation between statements I've made and these observations there's a definite reason why. The following were comments and observations I was given as part of a postmortem on a project I just wrapped up. They are customer comments, but I was given permission to relay them directly to whomever I felt might want to know: Background: They are sold on Tapestry - 3. They will not move to 4 unless there's a driving need due to monies expended to implement their current solution (they rolled their eyes at the mention of 5.0 stating the benefits didn't outweigh the costs) and lack of tooling. The "initial learning curve" was discounted as it was completely different from their previous framework and learning curves are everywhere - although I did point them to Kent's book and sold them on that (Thanks Kent!). The did feel this level of information should have been considered part of the overall tutorial/example documentation. Pros (for choosing Tapestry): 1. Tapestry allows separation of concerns 2. Component design allows different parts of the agency to have access to a library of components 3. Web interface designers do not need to know anything except HTML (conversely, Java developers do not need to know HTML) 4. Strong community support 5. An Apache project Cons with moving forward (to future Tapestry versions): 1. Codebase changes too radically between releases (this comment was made in discussions about 4.x and comments made about 5.0 - this brought up more conversations about "initial learning curve" with switching to a different framework having the same name) 2. Lack of documented reasons for change (they don't sharpen pencils without documented reasons - and before the jokes fly, yes - they do still use pencils :-)) 3. Too many ways to do the same thing (too flexible: w/ regard to Annotations vs. XML in 4.x conversations) 4. Lack of first-class tooling support within the framework (discussions on Spindle - which they love in T3 [thanks Geoff] and its lack of support in T4) 5. Perceived to be a "consultant's framework" in that it changes enough - and (back to the tooling) without tool support - that only consultants can keep up with how to migrate from version to version. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.5 (MingW32) iD8DBQFEafTkaCoPKRow/gARAs+yAKCyhzD0/y9GQIMwo6FYz/i1TTj3SACgtlP8 ghv9CBzskedeopIlbYGti4w= =DIjM -----END PGP SIGNATURE----- --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
-- Howard M. Lewis Ship Independent J2EE / Open-Source Java Consultant Creator and PMC Chair, Apache Tapestry Creator, Jakarta HiveMind Professional Tapestry training, mentoring, support and project work. http://howardlewisship.com --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
