Pascal makes a good point - that I think has been made in the past too - but perhaps a better way of looking at it is to make pyjs 'market driven'.
That is, let's not rush to build solutions to problems just because we can. We should (in the core pyjs project) focus on making pyjs more robust, usable, simpler to use/deploy, and perhaps most importantly, *extensible as a platform*. That's where interoperability issues might be addressed. I think ideally, pyjs should have well-defined architectural interfaces that others can use to build glue between pyjs and everything else as needed. I agree that pyjd is useful, especially as a development/debug platform, but the more interesting use the way the world is going now, is with pyjs web apps. IMO, the 'external' interface to pyjs in a web app is not well documented or accessible - at GBS (www.gbsware.com) our application plays a few tricks with frames and such to get a smooth integrated appearance that mixes Pyjs content and (in our case) Web2Py content, but we had to guess and fiddle about to get it right. Having an 'official' architecture for embedding and interacting with pyjs content would be a great advance. R On Sun, May 13, 2012 at 4:38 AM, Pascal Chambon <[email protected]> wrote: > Hi, > > Pyjd has always looked to me more as a "debugging helper" (to avoid > recompiling and loading JS, to have the robust tracebacks of native python) > than a production target. > With html5, cloud applications etc., the future seems to be the web, so imo > the best market would be web dev that dislike javascript and want to easily > code dynamic web apps (especially for backend admin tools, maybe, where > forms are far too limited). > > I'm pretty the biggest challenge of pyjs is integration with third party > libraries and frameworks. Eg. pyjs widgets should be able to call jquery's > excellent one-liners (like Sortable()) without hassle, or use misc. charting > libs without having to recode your own. > It would require proper value conversion on input and output, and some > warnings to pyjs users (using event handlers of several libs on the same DOM > elements is probably not a good idea), but it would more than certainly help > many people "cross the chasm". > > ++ > PKL > > > Le 13/05/2012 08:49, pca a écrit : > >> Hi, >> >> I read a great book about the promotion of disruptive technologies : >> "Crossing the chasm". It's been recommended to me by a friend working at >> Cisco : they use it extensively in their new product strategy. >> >> One of its main recommendation is to start by focusing on a niche market, >> and be absolutely the best at serving it. It stresses that it is a very >> hard choice, but a necessary one. I would suggest that pyjs is at a good >> point in time to make that hard choice. So, what is the core value >> proposition of pyjs ? Where is it absolutely the best solution ? Here are >> some possible answers, with concerns I have. I don't have a definite answer >> for this. >> >> * "improve your productivity by programming rich web client in python, not >> in javascript" >> but then, do I still have access to other javascript libraries, such as >> Dojo >> but do I really gain productivity when I have to compile my code, and >> when debugging the application is much more difficult ? >> >> * "create rich desktop application easily" >> but how about security issues at deployment on the desktop (local >> firewall, port conflicts) >> >> * "program once, for both web and desktop applications" >> but then, who needs that ? >> >> PC
