Verification

Dear students,

Here are some informations about the re-exam, if you are registered, that should come as a recall:

• The re-exam takes place this Tuesday 02.04.2019, from 10:00 to 12:00
• The write-up duration is 90min, so the actual examination time will be 10:15 to 11:45.
• The exams are open-book, which means you can bring any non-electronic document. This is due to the broad content of the lectures to be covered.
• Don't forget to bring an ID card or other official identification document.

Best regards and good luck!

-- 
Daniel STAN

EDIT: moved to room 301, two floors downstairs.

Dear students,

The results of Friday's exam are now published on the website, along with your grade.

We have scheduled an inspection next week Wednesday afternoon, from 14:00 to 16:00, in 528, E 1 3. If you cannot make it at this date, you can contact me to schedule another appointment.

 

Explanation about the results: because of the difficulty/length of the subject −sorry about this− we considered the last page (9points) of the last exercise, as bonus points, so passing the exam required 36 points instead of 45, over 90 points. 
If passing, each minitest was scaled to 5 points, and added as bonus to the total amount of points.

 

Best regards,

-- 
Daniel STAN

Dear students,

Here are some informations about the examinations:

• The first exam takes place this Friday, 22.02.2019, the re-exam on Tuesday 02.04.2019, from 10:00 to 12:00
• The write-up duration is 90min, so the actual examination time would be 10:15 to 11:45.
• The exams are open-book, which means you can bring any non-electronic document. This is due to the broad content of the lectures to be covered.
• Don't forget to bring an ID card or other official identification document.

 

Extra office hour: if need by some of you, I can offer an extra office hour in the coming week. A poll in the forum is available to pick a date.

Best regards,

-- 
Daniel STAN

Dear students,

An important remark was reported by some of you: the persistence checking asks to give up the search as soon as a terminal state is found. This assumption is made in order to reject models that have terminal states.

In the next question, the product of a system with a NBA of a LTL formula may introduce new terminal states ! The NBA is indeed not assumed to be total, and some trace fragments may not have a corresponding run in the NBA. Please construct in this question a product system that doesn't introduce new terminal states.

This can be implemented in the following ways (non-exhaustive list):
* Make the NBA total, by adding a sink non-accepting state (you can add a [true] guard from each states to the sink, since the NBA doesn't have to be deterministic);
* Add a sink state to the product LTS, whenever no other successor has been produced.

 

I hope this extra specification clarifies the subject and makes the last question easier.

Best,

--
Daniel