Yann wrote:I have been a RealBasic developer for many years, have tested the Web Edition several times and am buying a license right now. It's a very different product, not really suitable in its current state for creating big websites, but already impressive and increasingly powerful for prototyping and developing web apps in no time. I really wish RunRev would have taken that route for the LiveCode server, rather than their pretty anemic solution : full WYSIWYG editor for the controls, 100% RealBasic language for interface programing (not only for the logic and data management), standalone server built right from the editor, etc... but RunRev are far from that.
True, for database apps RB/WE is a nice toolkit.
For the LiveCode audience, however, it may not be an ideal fit, at least in terms of meeting expectations raised by the RevWeb plugin.
While a great many LiveCoders make DB frontends, there are at least as many making interactive graphical simulations, multimedia presentations, and other UIs that a RB/WE-like solution can't address.
The RevServer product takes us a step in that direction, allowing us to craft CGIs using LiveCode on the backend. In fact, I had the pleasure of using it to port a custom search engine we'd used in a desktop product to a web version, requiring only a few lines of additional code to wrap the output in HTML but otherwise it all just ran great without further modification.
While RB's being multi-threaded does indeed provide a measurable performance boost in allowing it to run gracefully under Fast-CGI, the nature of multi-threaded systems can require careful coding to avoid race conditions. The RevServer engine is single-threaded, but in practice the performance difference is relatively minor (best measured in milliseconds), and being a single-threaded CGI means that each instance is a discrete process, eliminating the potential for race conditions for the relatively low cost of a couple MBs of RAM.
True, that additional memory overhead can add up on servers with a lot of traffic, but few of us are making Google.
For most of the sites of the scope we manage, RevServer performs admirably well, sometimes surprisingly so. I've made CMSes with it that are not only simpler to use than Drupal, but require 1/5th the memory and 1/20th the processing time to generate a page. Given how many thousands of devs find the resource footprint of Drupal acceptable, besting it several times over should suffice for many real-world needs.
The biggest advantage RB/WE has over LC is the integration of Web UI layout in the IDE. They do a fine job with that, and while this is the sort of thing anyone in the LiveCode community could do as a plugin if they choose, apparently the community hasn't yet found the need great enough to do so.
I've written LC->HTML translators so I have a feel for what's involved. Mine have thus far been very specialized for specific apps, and indeed it's quite a bit of work to generalize for other uses.
But as far back as '97, ToolBook provided such tools to publish to the Web, so it's not like it's a new concept or somehow impossible. Indeed, CSS has only gotten more useful since then, so the opportunity is more ripe than ever.
I keep threatening to generalize my own translation tools for this, but alas my clients have kept me far too busy to pursue it. I've been hoping some FOSS-minded soul in the LC commnity will take an interest in managing such a project; I may be able to squeeze in enough time to contribute bits here and there.
In the meantime, HTML/CSS and JavaScript aren't particularly difficult to learn, and with things like the RevIgniter framework on the backend developers can at least enjoy a workflow that exceeds the productivity possible with most PHP-base toolkits, using stuff available right now. Not quite as fluid as RB's, but far broader in terms of the types of Web apps one can deliver.
Yann wrote:I'm not really happy to write this here on the LiveCode forums, because I had sincere hope in LiveCode on the web.
Personally, I'm very glad these forums support a diverse community with a broad range of experience, so I think your post was valuable and look forward to more from you. Diverse experience is how we all learn, and provides opportunities to encourage growth for the LiveCode platform in interesting and valuable directions.