Ticket #939 (closed planned_task: obsolete)

Opened 14 years ago

Last modified 13 years ago

SMOOTH_TEXT_RENDERING_R0

Reported by: Astea Owned by: vlado
Priority: 3 Milestone: M04_PRE4
Component: TEXT_RENDERING Version: 2.0
Keywords: Cc:
Category: MAIN Effort: 2.5
Importance: 0 Ticket_group:
Estimated Number of Hours: Add Hours to Ticket:
Billable?: Total Hours:
Analysis_owners: vlado Design_owners: vlado
Imp._owners: vlado Test_owners:
Analysis_reviewers: dido Changelog:
Design_reviewers: pap Imp._reviewers: meddle
Test_reviewers: Analysis_score: 3.5
Design_score: 3.5 Imp._score: 4
Test_score: 0

Description

wiki page: SMOOTH_TEXT_RENDERING_R0 - effort: 2.5d

Change History

comment:1 Changed 14 years ago by dido

  • Status changed from new to s1a_analysis_started
  • Analysis_score set to 0
  • Test_score set to 0
  • Design_score set to 0
  • Owner changed from Astea to dido
  • Imp._score set to 0
  • Analysis_owners set to dido

comment:2 Changed 14 years ago by dido

  • Status changed from s1a_analysis_started to s1b_analysis_finished

comment:3 Changed 14 years ago by deyan

  • Status changed from s1b_analysis_finished to new

Batch update from file query-1.csv

comment:4 Changed 13 years ago by vlado

  • Owner changed from dido to vlado
  • Status changed from new to s1a_analysis_started
  • Analysis_owners changed from dido to dido, vlado

comment:5 Changed 13 years ago by vlado

  • Status changed from s1a_analysis_started to s1b_analysis_finished

comment:6 Changed 13 years ago by vlado

  • Analysis_owners changed from dido, vlado to vlado

comment:7 Changed 13 years ago by dido

  • Status changed from s1b_analysis_finished to s1c_analysis_ok
  • Analysis_reviewers set to dido
  • Analysis_score changed from 0 to 3.5

The implementation idea section is not clear enough. So I'll summarize this.

  • First of all you give points to the tasks so they could be easily priorities.
  • Hardness and Expected effectiveness are the main aspects of an issue. Like cost and value.
  • The lower the summed number is, the higher is the priority of this issue.

So here are the attacked issues in order of their priority:

  • Improve logging(3)
  • Implement the issues commented as "performance"(5)
  • Memoization (part of a path, edge, etc.)(5)
  • Check for existing useless updates and reflowing (e.g. selecting the frame with highest z-order)(5)
  • Check for large ammounts of objects cloning(5)
  • Improve badness calculation(6)
  • Refactor badness(7)

Analysis review 3.5p (1h)

comment:8 Changed 13 years ago by vlado

  • Design_owners set to vlado
  • Status changed from s1c_analysis_ok to s2a_design_started

comment:9 Changed 13 years ago by vlado

  • Status changed from s2a_design_started to s2b_design_finished

comment:10 Changed 13 years ago by pap

  • Status changed from s2b_design_finished to s2c_design_ok
  • Design_score changed from 0 to 3.5
  • Design_reviewers set to pap
  • Generally I like the design and its nice structure but I'll leave you some remarks to look at.
  • It may be better to log the vertices with trace priority. If you take a look at http://sophie2.org/trac/wiki/LOGGING you'll see it is the level used to track the progress of a method.
  • Same goes for edges in the ther method. But as we have some problems with the logging system you may also leave them as debug, but make the inital loggings(the one that say the method is called) to info level.
  • The idea you express about changes is ok. GROUP_CHANGES_R0 is mostly in trunk as far as I know but there is much to do about them.
  • The thing about page element selection is quite hacky but I think I can accept it, because it is reasonable.
  • I think that if we refactor the HotTextSceneElement to be build like the other scene elements(having a DefaultHotTextSceneElement in the scenes module that has a synchronizing resource property) may improve things as it may change the layout not so often. I will be very glad if you do so.
  • I think the idea about AtomKinds is nice but I find this deep nesting a bit confusing. Maybe you can make the AtomKind and its subclases with default visibility and add them in the package of atom(maybe create some subpackage). This is not mandatory.
  • When I accept your design, it is not nice to change it anymore. Instead you should mention the differences in the Implementation section.
  • I'm a bit sorry that you couldn't make it to the most effective according to the analysis improvement - the memoization.
  • Good luck.

comment:11 Changed 13 years ago by vlado

  • Status changed from s2c_design_ok to s3a_implementation_started
  • Imp._owners set to vlado

comment:12 Changed 13 years ago by vlado

  • Status changed from s3a_implementation_started to s3b_implementation_finished

comment:13 Changed 13 years ago by meddle

  • Status changed from s3b_implementation_finished to s3c_implementation_ok
  • Imp._score changed from 0 to 4
  • Imp._reviewers set to meddle

I'm think that is a really good improvement and thanks to the fast fix with the cursor. The was not so bad like quality.

4p (90m)

comment:14 Changed 13 years ago by deyan

  • Status changed from s3c_implementation_ok to closed
  • Resolution set to obsolete

Batch update from file query-obsoleted.csv

Note: See TracTickets for help on using tickets.