• ThesisLog

Thesis Writing Log

This page tracks the status of my thesis writing. An updated outline of thesis appears here. Here's an initial dump of the coarse-grained revision history since the beginning of this document in mid-2006:

1. July 7, 2006: Created document to collect outlines, design docs, case study plan
2. July 10, 2006: Preliminary outline reformed into proper document with headings, collected content will be slotted into appropriate subsections; issues section added

3. July 12, 2006: Pasted architectural diagrams of Rainbow arch from AcmeStudio

4. July 24-5, 2006: Added TOC, expanded design details
5. August 3-5, 2006: Documented requirements traceability and case study needs

From this point on, each subsection represents a major revision.

July 2007

2007-08-02-ThesisArgument.pdf

• Added thesis arguments

August 25-30, 2007

2007-08-30-OutlineComplete.pdf

• Revamped outline

August 30-31, 2007

• Swapped Rainbow and Stitch chapter back to original order
• Worked on very beginning of intro and parts of 1.1
• Addition of "Evaluation" section in Discussions chapter, but refinement needed

2007-08-31-ChapterTalkShoe.pdf

• Minor edit to validation section on TalkShoe, and sent off section to Bob Pawlowski for review

September 1-2, 2007

2007-09-02-AddedRoadmaps.pdf

• Completed 1.4: This Thesis
• Worked on 2.5: Thesis Requirement, but not completed
• Started Chapter 3 and 4 intro paragraphs

September 3, 2007

2007-09-03-ApproachPartial.pdf

• Chapter 3: Approach in progress
• Drafted 3.1, 3.2.1, 3.2.2, 3.2.3, 3.2.4, 3.2.5
• Reordered sections

September 4-5, 2007

2007-09-05-Approach.pdf

• Completed draft of Ch 3
  • Moved requirement section from Chapter 2 to 3
  • Moved control loop discussion from Ch 3 to 6
• Separated conclusion chapter from Discussions
• Drafted abstract

September 6+8, 2007

• Ported SEAMS'06 paper for intro to Ch. 4, Ch. 4.1, Ch. 4.2
• Working on 4.3...

September 11-13, 2007

2007-09-13-StitchPartial.pdf

• Drafted 4.3
• Edits to 3.1 to put placeholders for architecture model and uncertainties
• Language formalism folded into 4.4

September 19, 2007

• Modification to thesis statement wording
• Met Bradley

September 21-22, 2007

2007-09-22-Planning.pdf

• Thesis end-game ponder and re-planning

September 24-25, 2007

2007-09-25-ApproachRevised.pdf 2007-09-25-LanguageIssues.pdf

• Completed 3.1 on fulfilling requirements, moving style subsection from 3.3 to 3.1, reordering subsections in 3.1
• Integrated style elements throughout framework component description
• Added some graphics
• Enumerated list of known and unsolved language issues

September 26-27 + October 1-3, 2007

2007-10-03-StitchAddedContent.pdf

• Reordered Rainbow and Stitch in Thesis Arg outline
• Enumerated Stitch issues, indicating what were resolved vs. unresolved
• Generated table of Stitch features and degree of need for semantic spec
• Completed 4.4.1 on simplifying assumptions

• Completed 4.4.3 and 4.4.4 with more fleshed-out specifications, addressing most of the s

• Added 4.6 on Stitch walkthru

• Missing: operational semantics, account for timing and control

October 4-5, 2007

• Paused writing, helped resolve AcmeStudio deployment issues

October 8-9, 2007

• Working on language semantics

October 10-15, 2007

• Revised and submitted camera-ready workshop paper for Int'l Workshop for Living with Uncertainties.

October 16-25, 2007

• Read "Dynamic Programming and Markov Processes"
• Worked out MDP example for Z.com with large, discretized states
• Developed small Perl script to solve MDP processes, and to run a bit of simulation
• Failed 1st attempt to convince David why the Strategy Selection problem requires MDP

October 26-29, 2007

• Prepared slides for SSSG talk on "Control Blues", and gave the talk
• Started composing slides for IWLU and Drexel
• Thought about what aspects of the Strategy Selection program required MDP; Monday meeting with David got him onboard to look at the MDP example for analytical benefits.

October 30-November 3, 2007

• Continued preparing slides for IWLU
• Refined MDP example and discussed example with David
• Met Mario of Coimbra and Peter of Kent

November 4-6, 2007

• Attended Int'l Workshop on Living with Uncertainties at ASE'07 in Atlanta
• Gave talk on where and how uncertainty is addressed in Rainbow
• Attended Steve Easterbrook's tutorial on empirical software engineering research

November 7-9, 2007

• IWLU expense and trip reports
• 8th: discussed thesis next steps with David and identified three important "holes" for thesis completion

November 10-17, 2007

• Worked on "thesis hole" slides:
  • Thesis evaluation strategy
  • Language design issues
  • Environment modeling plan
• Worked on Drexel guest lecture slides
• 13th: met with Vahe to go over MDP denotation mapping plans as well as future work on MDP-based analysis
• 15th: met with Jeff Barnes for initial chat on SoA; Marwan thesis proposal dry-run

November 18-20, 2007

• Traveled to Drexel, touched up guest lecture slides
• Discussed research collaboration possibility with Peppo
• Gave a guest lecture on software-architecture based self-adaptation in Peppo's SW Arch class

November 27-30, 2007

• Worked on book chapter of autonomic and networked systems
• Reported and liaised research collaborations with Peppo at Drexel
• Spoke with David about thesis time-line (and got OK) + future research (and potential after-grad employment)
  • next week: update committee members and run eval strategy by them to get them on-board
  • Dec-Jan: iterate thesis chapters with David
  • early-Feb: release full draft to committee members
  • mid-Mar: thesis defense

December 1-2, 2007

• Worked on Thesis Evaluation Plan slides, and sent off to David and Bradley for review
  • Added thesis status slides
  • Updated back-of-envelope task effort estimate
  • Added slide on plan to eval expressiveness
• Feedback from David, slides revised and completed
• Emails requesting for meeting sent off to committee members
• Resumed working on dissertation: Introduction chapter

December 3-5, 2007

2007-12-03-ThesisEvalPlan.ppt 2007-12-04-ThesisHoles.ppt 2007-12-03-TaskBasedAnalysis.xls 2007-12-05-IntroPartial.pdf

• Discussed thesis evaluation plan with Jeff Magee at Imperial College, and he's OK with it
  • Cautioned to avoid the pitfall of showing semantic equivalence to admin scripts
  • Noted "availability" as a better QoA dim than security
  • Favored coarse-grained task estimation (unless I have substantive evidence)
• Performed 1st-approximate effort of detailed task-based analysis
• OK'd thesis evaluation slides with David
• Scheduled thesis-evaluation discussion meetings with Jonathan and Peter
• Thesis Holes: slides completed for four holes, corresponding to four components of the Rainbow architecture
• Firefighting to restore waterfall.able from webserver compromise (mantis + bibAdmin)
• Thesis Introduction chapter
  • Composed 1.1 with 1.1.1-1.1.3 and 1.2
  • To complete: 1.3 on on limitations of current state-of-art, 1.3.1-1.3.3

December 6, 2007

2007-12-06-Introduction.pdf

• Introduction chapter (1) fully drafted
• To Do:

  • Revisit prose for Ch. 3 & 4

  • Revisit TalkShoe and other case study sections in Ch 6

December 7, 2007

• Revised archt font face tex-code, moved some discussion content from 4 to 7
• Revisited Ch. 3-4
  • Moved workflow section from 3 to 4
  • Weaved new MDP content into Ch. 4

December 8, 2007

2007-12-08-IntroApproachStitch.pdf

• Completed Black Friday statement, need to send to David
• Thesis Document
  • Added Evaluation Plan section 3.1.4 to Ch. 3
  • Added operational semantics, section 4.4.6
  • Added text to failure handling + preemption, section 4.3.5

December 9, 2007

• Composed detailed outline of IBM AWB Practicum report, yay!
• Sent revised Black Friday (updated practicum status) and practicum outline to David and Bradley
• Thesis Document (12-08 version revised)
  • Finalized abstract
  • Shuffled notes to appendix for more accurate content page count
  • Detailed edits of Ch. 1, 3, 4

December 10, 2007

• Thesis Document
  • Sent copy of Ch. 1, 3, 4 to David
  • Re-adjusted thesis schedules (last edit was September!)
• Met with Jonathan, who is OK with the evaluation strategy, including the case studies accomplished, but cautioned me on three elements:
  1. Think carefully about what is claimed with "user study"; talk with Bonnie John about methodology for user study
  2. When capturing admin script, be clear about boundary between Stitch script and admin script, needs enough detail to be convincing
  3. For task-based analysis, industry (IBM) data is quite crucial

December 11, 2007

• Thesis Document
  • Worked slightly on Ch. 6 exploratory portion
  • Worked on Ch. 5 up to 5.2.3
  • Generated Acme representation graphics for 3 Rainbow components
• Some more waterfall firefighting

• Contacted CMU network architect, Russ, who forwarded my request on to Dan McCarriar

• To do this week:
  • Complete Ch. 5
  • Compose Ch. 6, all but Z.com (75%)
  • Get at least 50% on Ch. 7

December 12, 2007

• Thesis Document
  • Worked a little on Ch. 5 up to 5.2.3

• Met with Bob at TalkShoe to go over case study report feedback, as well as chat about his assessment of adaptation development effort

  • Bob indicated roughly double of effort if adaptation capability developed without Rainbow
  • Bob believes that Rainbow adds maintainability, greater extensibility, and potentially lots of usability, if "sugar coated" with auto-deployment

December 13-16, 2007

2007-12-16-FrameworkEngineering.pdf

• Ch. 5 drafted
  • Revisited 5.1 to insert Design Qualities (moved from bottom of chapter)
  • Integrated probe/gauge protocols section into 5.2.2, translation layer
  • Worked on 5.2.4, 5.2.5, 5.2.6
  • Added 5.2.7 on Rainbow Deployment Architecture
  • Excerpted Z.com customization examples for 5.2.8, added customization detail in each section
  • Placed full Z.com examples in Appendix B
  • Completed 5.3, renamed Rainbow Adaptation Integrated Development Environment
  • Gathered sloccount data, completed 5.4

December 17-18, 2007

• Thesis Document
  • Folded in evaluation prose for Ch 7

  • Ch. 6, thesis claims + validation items, reordered Z.com to come after TalkShoe

  • Worked on detailed content for the first two (6.1.1, 6.1.2) system instantiations from 2004 IEEE article + (6.1.3) security instance

• Met with Peter to discuss thesis evaluation plan; Peter is OK with plan, but expressed 2 concerns:

  1. Cost-effectiveness: Qualitative task-based analysis looks like a workable approach, but sees industry data as adding credibility... factor of 2 is not significant, we're looking for 5-10x
  2. Composability: sees issue with generality of claim from 4 data points to N (all dimensions)... sees this as a fundamental problem of quality "tradeoffs", not a problem particular to the Rainbow approach, but Peter cautions me to be cautious about over-claiming
  3. Schedule restriction: can't schedule on Spring Break; in conjunction with Magee's preference, that leaves March 17-18th
• Correspondence established with CMU network admin to acquire admin scripts; will need to provide more research details

December 19-21, 2007

• Received David's extensive feedback on Ch. 1, 3, 4... lots of work ahead
• Met Bradley to discuss thesis-writing
• Thesis Document
  • Renamed Ch. 6, 7, 8 heading
  • Added summary section for Ch. 3, 4, 5
  • Shortened some paragraphs for layout
  • Began editing document based on David's feedback

December 21-27, 2007

• Organized thoughts on problem and solution, as well as traceability from claims to requirements to solution
• Christmas break

December 28, 2007

• Met with Bradley to discuss Ch. 4, ended up sketching out the threads from thesis hypotheses to requirements for the language (vs. requirements for the overall approach)

December 29, 2007-January 6, 2008

2008-01-04-IntroRevamped.pdf

• Thesis Document
  • Edits to document based on David's feedback: took notes of major issues in Ch. 1; added subsections 1.4.1 and 1.4.2 for contribution and roadmap
  • Populated thesis contribution content (1.4.1, 8.2)
  • Revised 1.3 and 1.4 to address David's significant feedback items; completed Ch.1 revision

  • Worked on hypotheses -> requirements traceability

  • Revised intro to 3; revised 3.1 (still need to revise 3.1.3), 3.2

  • Revised 3.3 to explicitly identify prior work, to offer explanations why; 3.4 on adaptation workflow moved back from Ch. 4

• Couple days of New Year celebration

January 7-10, 2008

2008-01-10-ApproachRevamped.pdf

• Monday: Met and spoke with David, briefly Skyped with Bradley
• Thursday: Met with David and Bradley to get Ch. 1 comments; David said Ch. 3 structure looks much improved
• Thesis Document (got linux lyx 1.4.x to work and generate PDF with change tracking)
  • Re-formatted 1.4.2 contributions into list form;
  • Revamped example in Ch.3 through 3.3
  • Revised 3.1.3 on utility theory
  • Weaved motivation + prior work through 3.3, 3.3.1-3.3.5
  • Added 3.3.6 complete example with Znn.com
  • Reviewed 3.4 on adaptation workflow
  • Ch. 3 completed and sent to David and Bradley
  • Outlined Ch. 4

January 11-19, 2008

2008-01-19-StitchNFramework.pdf

• Read up on dynamic semantics to help with later sections

• Discussed with Vahe feedback on 4.2 inputs/outputs (as well as ideas on language semantics --> Java snippets)

• Thesis Document
  • Minor improvements of argument in Ch. 1
  • Wrote Ch. 4 intro arguments
  • Completion 4.1 requirements + concept summarization table
  • 4.2 lang context, incorporated style and distinguish static vs dynamic inputs
  • Organized Stitch grammar, found bug due to lack of explicit operator definition
  • Introduced new 4.3.1 on adaptation objectives and conditions
  • Completed 4.3.2 operator, removed unnecessary mention of "atomic"
  • Completed 4.3.3 (tactic) to strengthen, walk thru example, explain model-atomicity
  • Completed 4.3.4 (strategy)
  • Completed 4.3.5 on trade-off
  • Added principle of what-when-how
  • Edited strategy grammar to reflect notion of "inline condition" (no change yet to reflect timer on action rather than condition)
  • Revised 4.5 Stitch walkthru for consistency with content, modified sub-section headings
  • Revised semantic traceability table
  • Recorded initial thoughts on how to handle Ch. 4 semantics section
  • Revisited Ch. 5 to improve intro and add traceability to requirements
  • Proofread chapter 5 text.
  • Accepted changes in Ch. 5
• Submitted new draft to David and Bradley, w/ Ch. 3, 4, and 5

January 20-22, 2008

• Took a break home

January 23-24, 2008

2008-01-24-IntroImproved.pdf

• Received Ch. 3 feedback from Bradley
• Thesis Document
  • Thought about/worked on 1.3.1 and 1.3.2

  • Slight rework of 1.4, should we term the 3rd hypothesis adaptability??

  • Began rework of Ch. 6 intro part

January 25-27, 2008

2008-01-27-ValidationProgress.pdf

• Thesis Document
  • Worked on Ch. 6, raised headings of experimental instances up to section-level
  • Reworked chapter intro
  • Completed 6.1 on C-S system
  • Wired up 6.2 on Libra
  • Composed 6.3 on University grade sys, adding new content on security operators and strategies
  • Adjusted strategy script due to bug in threshold comparison
  • Touched up 6.4 with validation limitations, noted that discussions need to go to Ch. 7, added summary
  • Added RAIDE mockup figure
  • Started constructing Ch. 7 evaluation sections
  • Uncommented list of figures and tables
  • Uncommented proposal-prose in Ch.2 to reveal all bibs and get an idea of how many
• Corrected encoding issue in bibtex file

January 28-February 1, 2008

2008-01-31-ValidationNSemanticsProgress.pdf 2008-02-01-EvalDiscussionRestructure.pdf

• Received feedback from Bradley on Ch. 4 requirements flow
• Received and noted Ch. 3 feedback from David
• Continued discussing lang semantics with David
• Thesis Document
  • Created thesis thread traceability matrix, enumerated preliminary set of discussion items
  • Began composing 6.5 on Znn.com, fleshed out 6.5.1-6.5.4
  • Minor grammar notational correction (::=)
  • Folded MDP + strategy-selection semantics content back into Ch. 4
  • Added some section labeling to allow traceability matrix references
  • Added Znn.com figures
  • Accepted all tracked changes in Ch. 5 and all prior tracked changes in Ch. 6

  • Worked on list of discussions in Ch. 7, dispersed Ch. 6 TalkShoe discussion content into Ch. 7

  • Created subsection structure of evaluation sections
  • Added new Ch. 6 files ahead of splitting evaluation and discussion chapters, split Ch. 7 into these two chapters
  • Moved Ch. 9 (formerly 8) thesis contribution from 9.2 to 9.1, slightly reformatted contributions list
  • Revisited Ch. 9 list of future work items, incorporating thoughts from slide on "enhancements"
  • Added draft versions from 1.27 and 1.31
• Fixed some BibTeX bug and dangling citation and figure references
• Added bibtex entries from Web articles

February 2-14, 2008

2008-02-14-IntroApproachRevised.pdf

• Noted detailed thoughts about operator, tactic, and strategy semantics
• Created slides on semantics, environment model, failure handling
• Met Bradley (experimental setup; hypotheses!) and David (language)
• Worked on summary table of Rainbow puzzle pieces
• Met David to continue discussing language issues and decisions
• Fixed/debugged Stitch grammar and interpreter
• Rewrote Znn strategies in updated grammar
• Debugged and tested strategies using Znn sim
• Committed changes to Rainbow impl + built 1.2.0 toolkit
• Thesis Document
  • Pulled content from Ali's report on sys-admin interview for 6.6, Bob interview for 6.7

  • Revised Ch. 3 revision, incorporate D&B feedback

  • Improved environment model discussion
  • Shortened text to fit in 16 pages
• Met David, good discussion, and submitted Ch. 1 + 3 to David
• Added Darwin bib reference
• Created draft thesis versions

February 15-17, 2008

2008-02-15-IntroApproachRevised.pdf

• Thesis Document
  • Merged prior changes in Ch. 3.
  • Begin rewriting Stitch chapter 4
  • About to move paragraphs around to reorder Context and Requirements sections.
  • Dramatically shortened intro
  • Swapped context and requirements sections,
  • Added Rainbow adaptation process as source of requirements
  • Reworked sys-admin illustration
  • Fixed grammar figure
  • Rewrote most of operator/tactic/strategy

February 18, 2008

2008-02-18-StitchRewritten.pdf

• Thesis Document
  • Rewrote 4.3.6 failure-handling
  • Revised and tightened 4.4 semantics, leaving an explicit placeholder for the operational semantics with list of evaluation rules
  • Added discussion tidbits
  • Added draft version

February 19-20, 2008

2008-02-20-FrameworkRevised.pdf 2008-02-20-StitchNFramework.pdf

• Thesis Document
  • Revised Ch. 5, moved framework arch types to new appendix A
  • Revised intro (moved req't intro to Ch. 7)
  • Shortened front Rainbow arch model portion
  • Dispersed customization contents across the Rainbow component sections and tightened text
  • Checked arch model/op-map

  • Revised AcmeGauge image to show only model w/ gauges

  • Added who-does-what-how table and weaved adaptation engineer role into customization text
  • Completed appendices on Rainbow Acme arch, Stitch, and customizations incl. Acme desc.
  • Spell-checked 4+5 before submission to David and Bradley
  • Generated draft prints for Ch 4 + 5.
  • Moved text to discussion sections.
  • Created separate appendix file.

February 21-27, 2008

2008-02-27-ExamplesNEvidence.pdf 2008-02-27-ExamplesChapter.pdf 2008-02-27-7_8_9_Outlined.pdf

• Sent off Rainbow SDK 1.2.0 to Heather Goldsby
• David positive feedback on Ch. 4, more suggestions for improvement
• Thesis defense scheduling
• Thesis Document

  • Edit Ch. 6 on Examples, incorporated TalkShoe interview content

  • Added Znn.com hypotheses
  • Improved Slashdot effect supporting evidence (added to Appendix)
  • Added slashdot and Znn.com prelim graphs
  • Revised Acme model to remove gauge and unnecessary environment model details
  • Revised Stitch scripts for updated grammar
  • Incorporated sys-admin interview results from Ali Almossawi's independent study content
  • Analyzed CMU network admin's Perl scripts (netbwe) for representation in Stitch.
  • Composed section for netbwe analysis from notes, drew arch diagrams, translated Perl script into a few Stitch strategies + tactics
  • Added Perl contents to Appendix
  • Recorded Znn.com development data
• Added Bib entries for personal communications
• Added draft versions for Ch. 6 and 7+8+9 outlines for David
• Met David for feedback on Perl analysis
• Met Peter to chat about thesis review expectation