Re: [be] future bibledit

[Dan Dennison]

On Tue, Mar 9, 2010 at 6:12 AM, Teus Benschop <address@hidden> wrote:

PHP is so flexible that making a mess of it can easily be done. It
depends on the programmer being responsible enough to keep the matters
clean.

It doesn't sound like I'm going to be able to convince you that using PHP for this project is a bad idea, so please allow me to convince you to use Facebook's PHP compiler:

• http://developers.facebook.com/hiphop-php/

At least then it'll be a bit faster, and more importantly, have a code validation process. Try it, I think you'll love it!
 

desktop applications. It means that the whole code base is going to be
rewritten, taking out the many inconsistencies now in bibledit.

Some thoughts about a web-based Bible editor:

• Don't fall for second system syndrome. Not every feature request can be implemented. Some of them will truly have to wait until you have more full-time contributors.
  
• Make it fast. What makes Bibledit great is its speed. An online Bible editor should focus on having a low latency user response.
  
• Make it scale. It's easy to write a web application that uses a database. It's not quite as easy or fun to write a web application that won't melt down when it takes a high load. Avoid using database joins. If it all possible, make it stateless. Make it work with nginx and other lightweight servers.
  
• Have unittests. Nothing will confound a developer more than being afraid to add features because you're unsure what will break. Use PHPUnit, and add them as you go, not at the end of the project.
  

Finally, I implore you to please keep the existing Bibledit GTK+ code available in its own project by itself. Other maintainers, including myself, will be contributing to it for some time. We've found it to be useful, fast, and full of features that were not available until now in a Free software Bible Editor.

Thanks for all of your effort on GTK+ Bibledit. I look forward to a new web-based editor.

dan :)

Rants below:

Everyone has their own favorite toolkit, but for cross-platform support with offline capability, the Google Web Toolkit fits the bill. Something you may really like about GWT is its use of Model-View-Controller-Presenter. That is, you can use a similar structural methodology that you use now in GTK+ with a web application.

http://code.google.com/webtoolkit/overview.html
http://code.google.com/webtoolkit/doc/latest/tutorial/mvp-architecture.html

GWT also has an Eclipse plugin, so you don't even have to leave your current environment.

----

Finally, the biggest drawback to an HTTP-based Bible Editor without offline capability is its local size requirements. Having to run a web server, database server, and web browser is too much for our users right now. They don't have laptops that are that powerful yet. That said, they might in 5-10 years, and we'll be happy it exists then.