[Trac #128] Meta bug to discuss new image formats for bootmenu

[Zarro Boogs per Child]

#128: Meta bug to discuss new image formats for bootmenu
--------------------------+-------------------------------------------------
 Reporter:  JordanCrouse  |        Owner:  JordanCrouse
     Type:  enhancement   |       Status:  new         
 Priority:  low           |    Milestone:              
Component:  linuxbios     |   Resolution:              
 Keywords:                |  
--------------------------+-------------------------------------------------
Comment (by JordanCrouse):

 '''xfig'''

 David Woodhouse proposed using xfig.  (This is a good summary of xfig:
 http://homepage.usask.ca/~ijm451/fig/).  Obviously, as a vector drawing
 format, xfig file sizes will vary greatly according to the complexity and
 number of colors in the source image, so care would have to be taken there
 not to exceed the baseline value.  parsing the file won't be too bad - it
 shouldn't add any additional size to bootmenu.

 The drawing functions on the other hand, will add some size to bootmenu.
 At the very least, we will need to support line drawing and poly line
 drawing which will add a minimal amount of space.  To fully support xfig,
 we would need elipse, arch and spline drawing, which will be much more
 complicated, draw very slowly, and probably require bringing in -lm (which
 we don't need now).

 We're not sure how much xfig will gain us.  I tried a tool called potrace
 which converts from ppm to xfig, but it only understands monochrome
 images, so while the resulting xfig was very small, it didn't draw
 correctly.  I tried another tool called autotrace, but the resulting file
 was twice as large as the .ppm, it was overly complex, and it looked
 horrible.  To accurately test the size difference, we would need to have a
 monochrome version of the .ppm that we could run through potrace.

 My current vote is that unless we can create very simple monochrome xfigs
 that only require line drawing that moving to xfig won't buy us enough
 space to be worth it.  I can be convinced otherwise though, if we get some
 good comparisions to go by.

-- 
Ticket URL: <http://dev.laptop.org/ticket/128#comment:1>
One Laptop Per Child <http://laptop.org/>