Hi all,

Todd and I have recently been discussing the possibility of migrating Pivot to 
run on SWT rather than Java2D. There are a number of reasons this might be 
worthwhile:

- It would allow Pivot developers to reach a broader audience by building 
Eclipse plugins or RCP applications using Pivot.

- It would make it considerably easier to build Pivot development tools (e.g. a 
GUI builder) in Eclipse (we already build the platform itself in Eclipse).

- Technical advantages of SWT over AWT (for example, SWT uses native font 
rendering - inconsistent text antialiasing is a common complaint about Pivot 
apps, and Java2D in general).

- It may allow us to generate some additional corporate interest in the 
platform. Sun/Oracle seem intent on pursuing Swing and JavaFX, which tends to 
make Pivot appear as an also-ran in the Java2D camp. By shifting to SWT, Pivot 
could become more appealing to IBM and others in the Eclipse ecosystem.

Pivot is architected in such a way that the impact of migrating to SWT would be 
minimal to most applications. However, the level of effort to port Pivot's 
skins to SWT would be non-trivial, so this is not something we would look to 
undertake without sufficient interest from the Pivot (and Eclipse) community. 

One potential hurdle is the lack of an "SWT plugin" that would allow SWT 
applications to run in a web browser (similar to an applet). While such a 
plugin could undoubtedly be created, it would require a lot of effort and 
support from the SWT community, support which they don't currently seem likely 
to provide. 

However, the association with "applets" has, to date, been a bit of a 
hinderance to Pivot's adoption. While an SWT plugin wouldn't actually be an 
applet, it may be perceived that way. I am wondering if it might be in the best 
interest of the platform to simply move away from browser-based deployment and 
focus on desktop delivery. Most Pivot applications I am aware of seem to target 
the desktop already. Perhaps we should simply focus on defining Pivot as a tool 
for building rich, cross-platform desktop applications rather than trying to 
directly compete with the big guns in the RIA space (as well as HTML 5, which 
will most likely draw attention away from all RIA platforms)?

This would be a significant change, so I would very much like your input on 
this possible direction. Please let me know what you think.

Thanks,
Greg

Reply via email to