A few quick notes. Appcelerator language independence - it's not independent, you still have to learn and use appcelerator's proprietary web extension language IN ADDITION to any language you might choose for the backend, and of course the regular mix of HTML+JS. So in my opinion it just shifts language dependency into framework dependency. With web2py and most of the classic MVC frameworks, even if you switch, you're not that much in trouble depending how the 'other' framework is handling things (with good refactoring tools you might be just a few views away), while with appcelerator it's pretty much all or nothing.
HTML design - mixed emotions, but improving. Having TR, TD and other HTML *functions* in code seemed wrong to me when I first saw them in web2py after all the PHP templating 'html in code is the dark side' mantra, but I quickly realized this is not the same issue here. The controllers DO have to have a rough idea what the view will display and can manipulate data much easier than the view can (think TR.append()). Your designer guy should be doing styles in CSS and maybe adding JS goodies to the view. It's not his job to change how many columns are there in your table, which rows are headings, etc. His work should be as data independent as possible. So it's actually not that bad of a mix - the coder still retains the control of the *generic* data layout, but the designer can work on the appearance part on a conceptual, global level, which in turn makes both the coder and the designer more effective as they don't need to sync all that often and are less likely to trip over each other. Also, the 'basic' or default layout generation is good as the coder can quickly prototype with the default element renderers and then commit the SAME code to the designer - if this needs to change, you're probably giving the (wrong) data back anyway and the view is *not* the place to change that. Admittedly, the implementation of this in web2py has a few rough edges, but love-it-or-hate-it, I do think the approach does make sense. Javascript. It really was a pain to develop in, but nowadays it's actually not that bad. With firebug and test frameworks like test.anotherway it's pretty easy to track down problems early on, this is not the same as the little exclamation mark you got in raw IE5-6, which really was developer hell. Also, it's hard to avoid with the interactivity web2 apps require. jQuery/JS toolkits - I didn't think it through but I'm under the impression it would be better to do tighter integration with these (maybe in an as-toolkit-agnostic-as-possible way) with helper functions, akin to what's done with the HTML tags mentioned above than go the appcelerator way. This would probably be T3 land as many have noted, and pretty far off, it needs to be seen where Massimo will take T2 first. On Oct 30, 11:58 am, "Daniel Guryca" <[EMAIL PROTECTED]> wrote: > Hi, > > My 2 cents: > > I have been a Java web and Swing developer for 5 years. > I use great Enhydra + DODS (database layer) frameworks and have > created some huge applications with these technologies. > I used to develop in PHP and Perl too but it's been an old history. > Those are my experiences from a real production world. > > From my perspective - I LOVE java (mainly it's pure syntax and all > free libraries around it - apache's libraries for example). > It's a very sad fact that there is no good lightweight web framework > for java (Grails is a new alternative but it uses RUBY like syntax > which I really hate (Perl like syntax) - subjectively very unreadable > for a team of cooparating programmers). > > But I have just started to look at some new (lighter and faster) RAPID > development tools. > > So yesterday, I took a Python user's guide and went through it - hey > that was so easy (but still prefer Java :-). > You know there are lot of good python web frameworks (that's why I > choose python). > > Then I downloaded web2py manual ... and took me around 4 hours to walk > through it's 80 pages + create 3 apps :-) > Very intuitive, simple and pure. To be true I have to say that I don't > like it's old JSP style. You know there is no complete web design and > code separation because of it. Morover forms are created with a python > - all design should be made from a HTML perspective only. I know - all > can not be perfect :-) > > As for an Appcelerator: > 1. Like it's general form - can be used from almost any language > (Java, Php, Python, Ruby, ...) so you are not stuck in 1 language !! > 2. Designer sees what you see (almost). > 3. No Javascript needed - I hate javascript - for anyone who knows > full java - it is a real nightmare to debug this stupid thing. > Dislike: > It does not seem to me to be a secure way to validate inputs on a client's > side. > > But I agree that almost everything can be done with a JQuery too. > > Have a nice day guys. > > On Thu, Oct 30, 2008 at 1:51 AM, achipa <[EMAIL PROTECTED]> wrote: > > > What people might be misunderstanding is my motivation here... There > > are hardly black-and-white choices, and when an engineering/business > > decision has to be made, it's usually for a particular situation. > > While the tone might have been a bit harsh, I'm not saying > > appcelerator sucks and nobody should ever use it. It's just that based > > on my experience, it is useable (with caveats) for very limited > > applications. OpenLaszlo does almost exactly the same thing, but as > > it's from 5 years ago, it uses tech that was all the buzz for serious > > web apps back then - generates flash from a Java backend, and the > > flash does the dynamic stuff from XML files you can 'code'. > > Appcelerator just adapts this to the present day buzz (replaces flash > > with JS widgets, uses HTML+AJAX instead of flash reloading XML files > > for the client side logic, instead of LZX markup language you have web > > expression language, etc). So *IMO* nothing revolutionary there, still > > the same (questionable) concept, still looks cool, still hard to > > develop, still hard to maintain, still has a strong lock-in, still a > > lot of untransferrable knowledge, and that is why I don't see > > appcelerator succeed (on a large scale) where OpenLaszlo had failed. > > But then again, I'm still earning my first million bucks, so maybe > > it's just me who doesn't get it ;) > > > With regard to toolkits - yes, we should use anything nice that fits > > our paradigm, but don't shift a paradigm to fit the tool (which is > > exactly why I don't think much of Silverlight and kin). If there are > > parts of Appcelerator that could be used to improve web2py, Massimo > > should use them by all means, but know your stuff, make conceptual & > > engineering and not buzz choices. > > > Take for example, in my case, PHP. Nobody argues that there are a lot > > of excellent sites done in PHP, huge userbase, a lot of development, > > frameworks, etc, etc, and yet, I *feel* there is no way I'm going back > > to PHP. Why ? PHP was good for doing simple stuff quickly. All the > > libs in one place, everything is global, you could make a dynamic web > > page in 2 minutes, and that's what drove LAMP to success. On the high- > > end, there was no question that for business app, you would use Java. > > As time progressed, web sites got more complex, and PHP itself went > > the way of more complexity, object orientation, templates, frameworks, > > soon namespaces. And here is the catch - in the process of migrating > > upwards on the business ladder, it just became too unwieldy (for me). > > The simple sites PHP was good at simply do not exist any more, and all > > the cruft from the age when quick and dirty was all the rage makes it > > cumbersome for the large ones (although as we can see the inertia is > > pretty strong, so PHP is still in high demand). Enter web2py. Python > > is just so much more elegant and more suitable for development of both > > system and web applications that when teamed with a nice framework > > (who said web2py ? :) it becomes my new 'ideal choice' for modern > > small-to-medium scale web apps. Simple (enough) and Powerful (enough). > > After a month with Python I didn't want to loose time with Perl again. > > After a month with django and web2py I didn't want to loose time with > > PHP (with Zend, Cake and Seagull) again. Do you see where I'm going ? > > You can do anything in almost any of these tools and languages > > (appcelerator and openlaszlo included), it's just the question of > > effort (which translates to money in a business environment), and I > > feel the sweet spot for effort for today's web apps moving *away* from > > PHP and into the land of web2py and it's kin. > > > Of course, you might disagree with my points above, they are just > > opinions, anyone can (and should) have them. As for bias, I know I'm > > biased, I tend to get biased towards things that make my engineering > > sense purr and generally make my life as a programmer easier :) I'm > > sorry if I stepped on any toes, it wasn't my intention, just wanted to > > share a few thoughts, sorry again if it got too rant-ish. > > > On Oct 29, 9:41 pm, JorgeRpo <[EMAIL PROTECTED]> wrote: > >> Well I havent used appcelerator either. > > >> But I found your opinion very personal and subjective. I am not going > >> to argue your points at this time. > > >> You dont seem to be criticzicing appcelerator buth the logic behind > >> its existence. > > >> So you dont agree neither with MS Silverlight, Mozilla Prism, Google > >> GWT, and others. > > >> But those frameworks are here and its better to take advantage of them > >> than to ignore them. > > >> On Oct 29, 9:54 am, achipa <[EMAIL PROTECTED]> wrote: > > >> > <disclaimer> > >> > I have not used appcelerator in any real life projects and just went > >> > through the docs/tutorials/screencasts. I did use half a dozen web > >> > oriented languages, frameworks and templating systems in the past ten > >> > years. > >> > </disclaimer> > > >> > Hate to be the Yin of all topics, but I find appceletator to be a bit > >> > heavy on the buzz side. What I think is the worst error in the concept > >> > is the Web Expression Language part. It somehow feels super cool and > >> > super wrong to me at the same time. It reintroduces program logic in > >> > places people were fighting years to get it out of, and doing it in an > >> > unreadable form (as it pretends to be html for syntax purposes, which > >> > is worse than any JS toolkit). This means two more things IMO - *real* > >> > designers won't be able to touch it, and it will have a strong lock-in > >> > factor as a design/UI made in it cannot be ported to any other > >> > framework (which is why all the JS toolkits actually make sense as you > >> > can adapt them to any underlying backend). 5 years ago I'd be super > >> > impressed with appcelerator, but now it seems to me as an approach > >> > with a dark side potential. Conceptually, it feels a lot like > >> > (Open)Laszlo, also a super cool and simultaneously super wrong > >> > approach. > > >> > On Oct 29, 2:04 pm, mdipierro <[EMAIL PROTECTED]> wrote: > > >> > > Thank you for your comments. > > >> > > I like appcelerator. I am reserving time over the christmas break to > >> > > learn more about it. Something may happen. > > >> > > Massimo > > >> > > On Oct 29, 2:38 am, Dunsun <[EMAIL PROTECTED]> wrote: > > >> > > > Hi, > > >> > > > It would be great to integrate Appcelerator with web2py. > >> > > > It would give us a superiority over all other frameworks. > >> > > > Web2py is clearly the best framework for python and Appcelerator is > >> > > > easy and very efficient way to add RIA and SOA to web2py. > > >> > > > Any plans for the future ? > >> > > > Thank you. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "web2py Web Framework" group. To post to this group, send email to web2py@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/web2py?hl=en -~----------~----~----~----~------~----~------~--~---
