Thesis Defense -- Pittsburgh PA

Xenophon · 04-24-2008, 08:35 PM

My defense is Tuesday April 29th at Carnegie Mellon.

Its open to the public, so anyone who feels like showing up is welcome.

Wean Hall 5324 2PM-5PM (although it should be over by 4PM). Abstract follows:

Concurrent programming has proven to be difficult. One cause of this difficulty is that the relevant thread usage policy seldom appears either in documents or code comments. A second difficulty is that thread usage policy---even when it is known---imposes widespread consequences on the code to be written. Finally, finding and removing concurrency faults in existing code is hard.

This thesis introduces thread coloring, a language of discourse useful for concise expression of and reasoning about intended thread usage policies in a wide variety of code. Thread coloring addresses a range of concurrency issues---assuring single-thread access, identifying possibly-shared data regions and localizing knowledge about roles for threads---that have not previously been comprehensively addressed. Using this language, programmers can model design intent about relationships among the roles of threads with respect to segments of executable code and also with respect to shared state. Programmers formally link the model with their code by expressing the model as annotations in that code.

This thesis describes a prototype analysis tool, integrated into an integrated development environment, and its use in case studies to demonstrate that thread coloring is a feasible and practicable approach to expressing and understanding thread usage policies, including complex ones. The tool analyzes consistency between the expressed model and the as-written code, and notifies programmers of discrepancies between them. The case studies use published code to demonstrate that developers can express useful models, identify concurrency faults and assure policy compliance. The thesis includes a demonstration of scaling to a medium-sized program of 140KSLOC and a demonstration of the potential to scale to much larger programs and support composition among analysis results for seperately developed components. By limiting the problem scope to thread usage policy, the prototype implementation requires one hundred times fewer annotations than are needed for full functional correctness---6.3 annotations per KSLOC, potentially reduceable in future by another order of magnitude.

This thesis provides five primary contributions to software engineering. First, it provides a language that developers can use to express thread usage policies. Second, it provides a systematic way to improve code quality by assuring that as-written code complies with expressed thread usage policy. Third, it uses a new combination of preexisting techniques to reduce the effort required to express models to very low levels. Fourth, it demonstrates techniques that permit the analysis to operate on very large programs---millions of lines of code appear to be within reach. Finally, it demonstrates techniques that permit straightforward and reliable incremental recomputation of results after a program change.

Boy oh boy will I be glad when that's over!

Xenophon

zelda_pinwheel · 04-24-2008, 08:37 PM

i am nowhere near pittsburgh so i won't be in the audience, but good luck !! it sounds like an interesting subject with really good practical uses. let us know how it goes.

Patricia · 04-24-2008, 08:41 PM

And I'm in the UK, so won't be there. But good luck, Xenophon.

RWood · 04-24-2008, 09:24 PM

I smell a future text book here. Not a Broadway musical; but, a book and a possible future seminar series. Heck, Tom DeMarco ran for years on just simple Data Flow Diagrams.

Nate the great · 04-24-2008, 09:36 PM

question: I know KLOC is an acronym for thousand lines of code. What does the 'S' stand for in KSLOC?

Also, can I have a copy of your dissertation? It sounds like an interesting read.

NatCh · 04-24-2008, 10:40 PM

I'm rooting for you, Xenophon! My wife just passed her defense on Monday, so I have some idea what you're up against.

Knock 'em dead!

Xenophon · 04-24-2008, 10:41 PM

Quote:

Originally Posted by Nate the great

question: I know KLOC is an acronym for thousand lines of code. What does the 'S' stand for in KSLOC?

Also, can I have a copy of your dissertation? It sounds like an interesting read.

The academic types seem to like KSLOC for K Source Lines Of Code. Why they had to have Source in there is a mystery to me.

Xenophon

P.S. As for your joke, well... In a couple of weeks the entire thesis will go on line as a CMU Tech Report, so it'll be available for all the fans out there. Gluttons for punishment, if they actually read the whole thing.

Nate the great · 04-24-2008, 10:54 PM

Quote:

Originally Posted by Xenophon

The academic types seem to like KSLOC for K Source Lines Of Code. Why they had to have Source in there is a mystery to me.

Xenophon

P.S. As for your joke, well... In a couple of weeks the entire thesis will go on line as a CMU Tech Report, so it'll be available for all the fans out there. Gluttons for punishment, if they actually read the whole thing.

I wasn't joking. I just finished my BS in Computer Science, and as part of my MS I'm planning to do a project on multi-threading.

NatCh · 04-25-2008, 12:00 AM

The joke is his signature line: Nate was being unclear about what he was making clear was meant as a joke. Which is kinda funny if you think about it.

Xenophon · 04-30-2008, 11:38 PM

Just in case anyone cares about the outcome, it went exactly as expected: pass with changes. I now have nine days to make all the changes the committee has requested. I'll finish the conclusion chapter ('cause I didn't get it done before the defense) -- no surprise there. They also want me to re-work the theory chapter, which will be something of a challenge to get done before COB on the 9th of May.

Somewhat to my surprise, my outside member (Guy Steele -- the same one who's on the Java language reference manual; a Sun Fellow; and an ex-office-mate from very long ago) showed up in person for the defense. He was very positive about the whole thing, which left me walking on air.

After all, he's one of the very smartest people I know. So maybe if Guy thinks it's good stuff, I should believe it...

Xenophon, enjoying the first day I've taken off in two-and-a-half months.

RWood · 05-01-2008, 12:06 AM

Congratulations.

Remember to post a link to it when it is posted on line Mr. Doctor to be.

Indigo Ink · 05-01-2008, 12:19 AM

Quote:

Originally Posted by Xenophon

Xenophon, enjoying the first day I've taken off in two-and-a-half months.

Congratulations and best of luck for May 9th. I defended my thesis a few months ago and I know how exatcly it feels 'the day after' - exhausted

By the way, I hope you don't wear what your avatar wears when you go to defend your thesis do you

NatCh · 05-01-2008, 12:43 AM

Congratulations, Dr. Xenophon!

Sparrow · 05-01-2008, 01:15 AM

Brilliant news, great to hear it went well.

Many congratulations

zelda_pinwheel · 05-01-2008, 08:18 AM

yay for you !! bravo.

04-24-2008, 08:35 PM	#1
Xenophon curmudgeon Posts: 1,487 Karma: 5748190 Join Date: Jun 2006 Location: Redwood City, CA USA Device: Kobo Aura HD, (ex)nook, (ex)PRS-700, (ex)PRS-500	Thesis Defense -- Pittsburgh PA My defense is Tuesday April 29th at Carnegie Mellon. Its open to the public, so anyone who feels like showing up is welcome. Wean Hall 5324 2PM-5PM (although it should be over by 4PM). Abstract follows: Concurrent programming has proven to be difficult. One cause of this difficulty is that the relevant thread usage policy seldom appears either in documents or code comments. A second difficulty is that thread usage policy---even when it is known---imposes widespread consequences on the code to be written. Finally, finding and removing concurrency faults in existing code is hard. This thesis introduces thread coloring, a language of discourse useful for concise expression of and reasoning about intended thread usage policies in a wide variety of code. Thread coloring addresses a range of concurrency issues---assuring single-thread access, identifying possibly-shared data regions and localizing knowledge about roles for threads---that have not previously been comprehensively addressed. Using this language, programmers can model design intent about relationships among the roles of threads with respect to segments of executable code and also with respect to shared state. Programmers formally link the model with their code by expressing the model as annotations in that code. This thesis describes a prototype analysis tool, integrated into an integrated development environment, and its use in case studies to demonstrate that thread coloring is a feasible and practicable approach to expressing and understanding thread usage policies, including complex ones. The tool analyzes consistency between the expressed model and the as-written code, and notifies programmers of discrepancies between them. The case studies use published code to demonstrate that developers can express useful models, identify concurrency faults and assure policy compliance. The thesis includes a demonstration of scaling to a medium-sized program of 140KSLOC and a demonstration of the potential to scale to much larger programs and support composition among analysis results for seperately developed components. By limiting the problem scope to thread usage policy, the prototype implementation requires one hundred times fewer annotations than are needed for full functional correctness---6.3 annotations per KSLOC, potentially reduceable in future by another order of magnitude. This thesis provides five primary contributions to software engineering. First, it provides a language that developers can use to express thread usage policies. Second, it provides a systematic way to improve code quality by assuring that as-written code complies with expressed thread usage policy. Third, it uses a new combination of preexisting techniques to reduce the effort required to express models to very low levels. Fourth, it demonstrates techniques that permit the analysis to operate on very large programs---millions of lines of code appear to be within reach. Finally, it demonstrates techniques that permit straightforward and reliable incremental recomputation of results after a program change. Boy oh boy will I be glad when that's over! Xenophon

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
New member in Pittsburgh	hrosvit	Introduce Yourself	8	07-10-2010 12:11 AM
Hi from Pittsburgh	Ranagirl	Introduce Yourself	11	05-21-2010 12:48 PM
Hello from Pittsburgh PA	jrl007	Introduce Yourself	9	02-16-2010 05:49 PM
Hello from frigid Pittsburgh	King Windom	Introduce Yourself	17	01-03-2010 07:09 PM
Cal Poly University Senior Thesis Survey: How Society Will Read in the Future	dbluke	Lounge	11	05-23-2009 03:53 AM

04-24-2008, 08:37 PM	#2
zelda_pinwheel zeldinha zippy zeldissima Posts: 27,827 Karma: 921169 Join Date: Dec 2007 Location: Paris, France Device: eb1150 & is that a nook in her pocket, or she just happy to see you?	i am nowhere near pittsburgh so i won't be in the audience, but good luck !! it sounds like an interesting subject with really good practical uses. let us know how it goes.

04-24-2008, 08:41 PM	#3
Patricia Reader Posts: 11,504 Karma: 8720163 Join Date: May 2007 Location: South Wales, UK Device: Sony PRS-500, PRS-505, Asus EEEpc 4G	And I'm in the UK, so won't be there. But good luck, Xenophon.

04-24-2008, 09:24 PM	#4
RWood Technogeezer Posts: 7,233 Karma: 1601464 Join Date: Nov 2006 Location: Virginia, USA Device: Sony PRS-500	I smell a future text book here. Not a Broadway musical; but, a book and a possible future seminar series. Heck, Tom DeMarco ran for years on just simple Data Flow Diagrams.

04-24-2008, 09:36 PM	#5
Nate the great Sir Penguin of Edinburgh Posts: 12,375 Karma: 23555235 Join Date: Apr 2007 Location: DC Metro area Device: Shake a stick plus 1	question: I know KLOC is an acronym for thousand lines of code. What does the 'S' stand for in KSLOC? Also, can I have a copy of your dissertation? It sounds like an interesting read.

04-24-2008, 10:40 PM	#6
NatCh Gizmologist Posts: 11,615 Karma: 929550 Join Date: Jan 2006 Location: Republic of Texas Embassy at Jackson, TN Device: Pocketbook Touch HD3	I'm rooting for you, Xenophon! My wife just passed her defense on Monday, so I have some idea what you're up against. Knock 'em dead!

04-25-2008, 12:00 AM	#9
NatCh Gizmologist Posts: 11,615 Karma: 929550 Join Date: Jan 2006 Location: Republic of Texas Embassy at Jackson, TN Device: Pocketbook Touch HD3	The joke is his signature line: Nate was being unclear about what he was making clear was meant as a joke. Which is kinda funny if you think about it.

04-30-2008, 11:38 PM	#10
Xenophon curmudgeon Posts: 1,487 Karma: 5748190 Join Date: Jun 2006 Location: Redwood City, CA USA Device: Kobo Aura HD, (ex)nook, (ex)PRS-700, (ex)PRS-500	Just in case anyone cares about the outcome, it went exactly as expected: pass with changes. I now have nine days to make all the changes the committee has requested. I'll finish the conclusion chapter ('cause I didn't get it done before the defense) -- no surprise there. They also want me to re-work the theory chapter, which will be something of a challenge to get done before COB on the 9th of May. Somewhat to my surprise, my outside member (Guy Steele -- the same one who's on the Java language reference manual; a Sun Fellow; and an ex-office-mate from very long ago) showed up in person for the defense. He was very positive about the whole thing, which left me walking on air. After all, he's one of the very smartest people I know. So maybe if Guy thinks it's good stuff, I should believe it... Xenophon, enjoying the first day I've taken off in two-and-a-half months.

05-01-2008, 12:06 AM	#11
RWood Technogeezer Posts: 7,233 Karma: 1601464 Join Date: Nov 2006 Location: Virginia, USA Device: Sony PRS-500	Congratulations. Remember to post a link to it when it is posted on line Mr. Doctor to be.

05-01-2008, 12:43 AM	#13
NatCh Gizmologist Posts: 11,615 Karma: 929550 Join Date: Jan 2006 Location: Republic of Texas Embassy at Jackson, TN Device: Pocketbook Touch HD3	Congratulations, Dr. Xenophon!

05-01-2008, 01:15 AM	#14
Sparrow Wizard Posts: 4,395 Karma: 1358132 Join Date: Nov 2007 Location: UK Device: Palm TX, CyBook Gen3	Brilliant news, great to hear it went well. Many congratulations

05-01-2008, 08:18 AM	#15
zelda_pinwheel zeldinha zippy zeldissima Posts: 27,827 Karma: 921169 Join Date: Dec 2007 Location: Paris, France Device: eb1150 & is that a nook in her pocket, or she just happy to see you?	yay for you !! bravo.