Printing consensus poll

This page is outdated. It is kept for historical reasons, e.g. to document specific decisions in Inkscape development.

Problem

Bryce has pointed out the #1 feature need for Inkscape right now is strong support for Adobe Portable Document Format (PDF) and Postscript (PS) in his 2005.11.30 email to the Inkscape mailing lists. This ConsensusPoll exists to enable the community to efficiently create and commit to a plan of action to resolve this issue.

Resolution

This poll has not elicited sufficient interest / participation to move forward. We require 20 people list themselves in the poll, but as of Apr 30, 2006 only 17 people have signed up.

$300 is pledged, but that is insufficient to match the bids that have been made.

There is no concensus achieved as to what specific issues should be addressed.

The plan needs to be further developed, but only a few people have worked on it. Also, despite these efforts, most votes have remained 'no'.

Contract

• >= 20 inkscape enthusiasts
• >= 4 developers with commit access
• >= 95% yes
• all conditions in plan are met
• 72 hour GO timer
• 80% cloture

ContractExplanation

Because this contract defines what it means to participate in this poll it will never be changed. If it turns out to be inadequate, this poll must be canceled and a new poll with a new contract must be drafted.

Yes-No Poll

Participation in this poll is open to anyone who wants to participate.

Voting yes in this poll means that you believe the community created plan is ready for implementation, and that you will do your part to support its execution. Voting no in this poll means that you have outstanding concerns about the community created plan. You can change your vote at any time. All votes are transparent so that those voting yes can listen to the concerns of those voting no.

No ... User:BrandonCsSanders    - Success needs to be defined (e.g., if all 20 test cases are passed ...)
Yes... User:BryceHarrington     - Let's do it
No ... User:AlexandreProkoudine - There is no plan yet
No ... User:ColinMarquardt      - There is no plan yet
No ... User:EricWilhelm         - The plan needs salt
No ... User:MenTaLguY           - This isn't much of a plan...
No ... User:JonCruz             - Plan? Not quite yet
No ... User:JonPhillips         - no plan...
No ... User:NicuBuculei         - so far the plan is weak
No ... User:RalfWStephan        - Where are the test cases?
No ... User:AndyFitzsimon       - I will also pledge $50 once someone defines how we do pledges.
No ... User:TuukkaPasanen       - There isn't good plan yet.
No ... User:EricJonas           - There is no plan
No ... User:AdibTaraben theAdib - There is no plan
No ... Martin Owens        - There is a plan but it's pants.
Yes ....User:Corey Burger      -plan, what plan?
No ... User:Gustavb             - The plan is incomplete.

(If you are voting no because the plan is incomplete, please describe what you'd like to see added or changed.)

Pledges

$50 ... User:BryceHarrington  - I will pledge $50 once someone defines how we do pledges :-)
$50 ... User:AndyFitzsimon    - I will also pledge $50 once someone defines how we do pledges.
€50 ... User:RalfWStephan     - whenever the plan has succeeded
$50 ... User:BrandonCsSanders - once the plan is done
$100 ... User:EricJonas       - Just let me know how to pledge

Plan

1. Describe the Current PDF Support and the known PDF bugs
2. Identify Required PDF Support
3. Assemble a few dozen examples to use as PDF test cases
4. Document how each of the current PDF tools (potential starting points) performs on the suite of test cases
5. Determine timeframe
   1. Start some time in early 2006?
   2. Expected completion some time in mid/late 2006?
6. Select developer with the right skillset
   1. C/C++, XML
   2. PDF file format
   3. SVG file format
   4. Able to work under (U.S.) contract
   5. /Other skillset requirements?/
7. Determine what funding is needed to enable them to complete the work
8. Establish a contract with the selected developer to perform the work
9. When the developer is done, verify that all test cases pass, then pay developer

Background Information

• SOC_Original_Project_Prompts
• PrintingSubsystem
• Roadmap (Milestone 11) Import/Export Feature Enhancements

Adequate Technical Solutions

What solutions would be fine?

• Library that integrates directly in to inkscape (and scribus, ...)
  • Use Cairo as the presentation layer and use its support for pdf/ps
    • This may be a first step towards an eventual migration of the backend renderer to Cairo
  • See Poppler for PDF import.
    • pdftosvg 1
    • pdftosvg 2
    • pdftosvg 3
  • Pull out the scribus pdf support into a libpdf (libpdf name has already been used) that can be shared by both scribus and Inkscape
    • Requires conversion of Qt-isms into more neutral widgetset-independent style
• Extend an existing standalone (filter) tool that can be bundled with inkscape
  • VectorSection
  • pdf2svg
  • ill2svg
  • ai2svg
  • pdftk

Possible Strategies

• Raise money and hire a developer to implement it
  • Collect pledges, once we are over threshold collect funds (http://fundable.org/)
    • What should the threshhold be? $2000? $5000?
    • Still too early to decide what the threshold should be. We need to know how much it will cost first.
    • Fundable now charges a 10% fee from the amount collected
  • Need to decide what the funds would be spent on ahead of time
    • 10-20% - Tester to collect test cases and put together a test suite for developer to use
    • 70-80% - Developer to write the code, document it, and make it pass the tests
    • 5-10% - Liaison to handle the paperwork, collect/distribute funds, track/report progress, and verify the work has been completed adequately
  • Mercenaries that would do this for us?

• • • User:BrandonCsSanders ... I'd do it for $2000, half up front, half once it meets spec. Caveat: it would take me six months to complete because my plate is pretty full right now and I'll be snatching a week here and a week there.
    • User:EricWilhelm ... This will be a lovely fit for VectorSection. I'm guessing $5k would get it to about the same level as XAR and SVG in chromista. I would also be happy to oversee/direct the work of someone else wanting to write this as a pair of VectorSection connectors (crs2pdf, pdf2crs.)
    • /other mercenaries?/
• Hold a BarnRaising
  • at RecentChangesCamp in Portland February 3-5
  • other location/time?
• Wait for a volunteer to get inspired and just do it

Questions/Brainstorms

• What about what inkscape earned through the Google Summer of Code? What's the status of this?
  • $2000 earned; there has been no plan or consensus for what to use it for
• Does Cairo have PDF export already? What about to use cairo for just this job? and later then use cairo for the whole presentation layer? Instead of creating an interims solution?
  • Yes, Cairo has PDF export capabilities; it's unknown if it's significantly better than our current (poor) pdf export functionality
  • Cairo Postscript and PDF backends are still experimental and disabled by default in current stable release, but plan is to make them supported in next stable release, expected for the end of 2005. See Cairo ROADMAP
  • A way to test cairo capabilities would be to use librsvg HEAD. See Dom Lachowicz blog entry.
• Investigate ps2ai, pstoedit and ai2svg conversion options (potential EPS support?)
• Import of native Adobe Illustrator files, which have been PDF based since Illustrator 10 (at least).
• Other Bugs and RFE's related to Import or Export?
• VectorSection
  • Other than not yet being stable, this seems like a very good choice and provides lots of choices for import/export.
• Is direct import/export of PS or PDF really necessary? Or is transparant import/export a reasonable target? The main problem with the current implementation is the need for the users to deal with filter dependancies. A complete bundle of the most important filters would solve this. Filters have the advantages of allowing Inkscape to concentrate on what it does best, SVG, and allowing sophisticated users to insert specialized filters into existing pipelines.