User Tools

Site Tools


Master project of Jelte Leijenaar

This is the page of the Master project of Jelte Leijenaar. This project is done for Neopost Technologies B.V. and this page is the main communication channel for this project.




This page is for questions and answers between the Project participants

First I would like to emphasize that you don't have to answer all the questions in one go if there isn't much time. Just pick the ones you think are important to answer first. The others will come next mini-deadline.

Question: 20-04-09

  • Q(J): To be able to make a decent planning I would like to know when you have planned your vacation this summer. Unfortunately I probably wont be able to graduate before the end of August, due to vacations of Bert van Beek and Prof. Rooda in July/August. At this moment I've planned to graduate within the first two weeks of September. Unless you are on vacation in that weeks.
  • A(D): I will have a leave in September, and I think it will be in the second half of September. There is an uncertainty present though. I will know details in the next three weeks (say before 15th of May).

Question: 17-04-09

  • Q(J): The error-handling I noticed when using the DS-86 is that when a document gets stuck, the machine is stopped and the user needs to remove the stuck document. Then the machine empties itself and waits for input from the user to restart the job. Is this correct?
  • A(D): I am sending you email with contact data of Erik Huizinga, which told me to be glad to answer the global behaviour kind of questions.
  • R(J): Please read the Progress page and place remarks about it on this page. If I pose a question on the progress page, I'll also post it here.
  • A(D): The general remark is that you add hronological updates, and not the updates in the form of your your final report. For me it is bellow an acceptable efficiency limit to follow your modelling progress stepwise. I indicated “you can proceed with your reporting work on the Wiki, but then open a separate page, and clearly indicate which material will go to which report chapter” - so please reshape the page in the form of sections. And indicate in the title of the section by which date you want my feedback.

Update: 09-04-09

  • R(J): Today we had a very fruitful discussion about the project. Instead of delivering a complete supervisor for the whole system, we're going to deliver a supervisor which has very useful properties. I already had interchangeability very high up my list, but its even higher now. I'm going to treat most the components (ie. Collator, Vertical Track, Folder) as separate processes, which have their own supervisor. To assure correct behaviour an interface will have to be made between these components. The events in the interface will be the only shared events in the components. Next week I will post a picture of the interface between the Vertical Track and the Collator to illustrate this concept.
  • R(J):This, however, does mean that I'm not able to update the report as early as the 1th of May. In exchange I'm going to put up pictures and small reports on this wiki every week. This is all in favour of the usefulness of the project for Neopost, so I hope you can agree on the above.
  • A(D): I can adjust the “1st of May” requirements:
  1. please send me the hardcopy of my last review back
  2. and the corrected version of the reviewed parts
  3. you can proceed with your reporting work on the Wiki, but then open a separate page, and clearly indicate which material will go to which report chapter.

Question: 27-03-09

  • Q(J): Since some of the subjects which I'm going to discuss in my report are probably confidential, I'd like to know this before I actually start writing about it. I'm going to give the layout of my report and some remarks about the contents. I would like to ask you to give your opinion about confidentiality of the chapters.
    • Summary
    1. Introduction
    2. Description of the DS-86: This is basically chapter 3 of the half-way report. (Probably confidential)
    3. Description of Supervisory Control Theory (this isn't confidential)
    4. Supervisors for the DS-86 (divided into Feeder Vertical, Interface and Inserter): This will include several Automata of control-loops of the DS-86. (Probably confidential)
    5. Simulation Results
    6. Conclusions and recommendations
    • Appendices: CIF code of the plant model and STS code of the supervisor (these are probably confidential)
  • A(D): You do not need to constraint your writing by any of the confidentiality issues, because when formulating the assignment I chose for a well known existing product. Anything you describe on the existing matters will not be confidential. Only parts of new generated value may need some scrutiny, and that will be the code of the generated supervisors, as you correctly noticed in your last bullet. But you will have to generate this anyway, so we will decide in the last stage how to present it.

So please go on writing full steam!

Question: 03-03-09

  • R(J): As you have already found out I've mailed you my Half-Way report and I would like to get some feedback about it.
  • A(D):
  • The comments are two kind: red remarks that require due corrections, and green and all not-coloured remarks pertaining to some style suggestions which made the text easier for reading to me
  • The language used has in general a very good level and style. The text is pleasantly readable. The abstractions in the text are proper.
  • Switching chapters 2 and 3 would improve the logic of the text flow and support understanding of the subject of present Chapter 2 after having introduced the machine functionality and components presented in present Chapter 3. Perhaps some parts of the chapters should be shuffled around (also maybe at the end of the Chapter 1).
  • The text misses a proper references treatment – perhaps this will be easier to do in the final touch upon the text, but be advised not to leave it for the last days.
  • As earlier indicated, putting captions on the machine parts drawings would dramatically improve readability and overall quality of the text.

Question: 16-02-09

  • R(J): You said last question round that you wanted to have a concept report including all what has been done since 7th of January. In that time I have been getting to know CIF and STS and the theory around it. I'm not going to be able to complain the theory in whole yet. But I'll try to explain everything I know. Please be aware that the report will be a very first concept. So a lot will change in the next writing round.
  • A(D): Caution! It says: “reporting on all what has been achieved by the time the presentation was given in Drachten on 7. January 2009” - that means: all what has been achieved since the beginning of the project till 7. January 2009. Therefore, I want your best effort towards the final report on all the topics before you started working on the STS theory for the controller synthesis. I do not require any reporting of the synthesis in your 2. March draft, as I am also not interested in thorough review of immature parts. If you however do, not the same quality of that part would be expected as it is on the “up-to-the-7th-January” part, which should be your best effort!
  • R(J): About CIF and STS, I have a simple model of the feeder running and I keep expanding it with pulse-driver and virtual sensors till that part is complete. The Inserter part should be more or less routine after that. The progress last weeks was a bit slow, but I'm gaining momentum.
  • A(D): That's certainly good news that you get that part up and running while putting significant time on the reporting on the previous results.

Question: 12-01-09

  • R(J): To be able to simulate my model with the State Tree Structures Supervisory Theory, I eventually need to rewrite my model to the CIF version of Chi. This will, offcourse, take some time during the deduction of the supervisor.
  • A(D): This is all okay provided a good version of the report concept, including reporting on all what has been achieved by the time the presentation was given in Drachten on 7. January 2009, is delivered for my review by 2. March 2009.
  • Q(J): How was the response after my presentation last week?
  • A(D): The reasoning about DS-86 as you presented it was experienced as interesting. The interesting is to observe which abstractions are used to understand the functioning of the system, and in what way the used modelling results present aspects of that functionality. The questions you had are observed more valuable than the model at this stage, because the questions you put in order to better model the system caused reasoning about some system aspects which had not got a proper treatment earlier - and that is always one of the important benefits of any modelling attempt. It is still not clear how Neopost will benefit of the models you are building - I expect this will get clearer when you end up with some control generation results.

Question: 18-11-08

  • Q(D): When do we appoint the next (and perhaps the last one in this year) presentation in Drachten, which would be a follow-up of the last TWINS-NL meeting presentation of 07-11-08?
  • A(J): I would like to combine that presentation with a meeting with Aalzen about the description of the DS-86 which I'm going to give to him this Friday. So I would like to give him a week or 2 to discus it with the other engineers. That would mean 10 or 11 December or one week later 17 December.

Questions: 30-10-08

  • Q(J): I will give a presentation about the hybrid-/chi models of de DS-86 so far. Can I use the full schematic overview of the DS-86 including the sensors and actuators? These overviews will dramaticly clarify the models.
  • A(D): Yes with no hesitation, especially for the TWINS audience.
  • R(J): I'm now occupied with typing the process descriptions of the DS-86, which I will send to Neopost begin December at the latest. For the next 2 weeks I won't produce any other products. Not besides the presentation of course, see you there.
  • R(D): Even better to come to present (if it involves the Chi code). This product is enough, don't worry :-)

Questions: 23-10-08

  • Q(D): With which information from your 22/10 visit to Neopost is your concept of the DS-86 now more complete? Which are still insufficient defined areas?
  • A(J): As far as I know I now have all information I need to make a model with the basic functionality of the DS-86. I know where I need to incorperate the Pulsedisks and with the overview I got from Joost with the distances I can calculate how much pulses these distances are. I'll make a description of the working of the DS-86 in text (which will be included in the final report) which he (and some other experts) will check. This because it would take to much time for Aalzen to understand the /chi model which already consist of almost 700 lines. The description in that document, with the remarks of Aalzen, will be how the final model works.
  • Q(J): Can we have a meeting on the 7th of November during or after the Twins meet?
  • A(D): Certainly I would go for it, only I have not yet got the agenda for that TWINS meeting.

Questions: 02-10-08

  • R(J): I added a Project Planning page. I think this will cover yours and Joost's request for a 'Stappenplan'. I also added some points after which I would like to come to Drachten to present some results. Please give remarks about this planning. (The Project google agenda will stay updated with my appointments regarding this Masters Project)
  • R(D): You get the feedback this week.
  • Q(J): During our meeting last week the follow-up to the DS-86 was mentioned. Do you, or Joost, want me to model that machine? This offcourse will take some time in getting to know the differences between the machines. If so, please tell me the differences between the machines so I can estimate how much more work has to be done.
  • A(D): The differences are lower than the level of detail you can model.
  • Q(J): Last Monday I recieved some short answers to my questions of last week, but there are still some questions open. When I want to continue modelling I'm going to need more information about the subprocesses.
  • A(D): Propose how to come to more information; I followed the link “subprocesses”, but what is specific to subprocesses on the page
  • R(J): I mean the exact processes of the Folding table or the Feeder/Vertical Track for example. I can always make a model the way I think it works and then get feedback later on.
  • Q(J): To be able to judge if my modelled system is working like the DS-86, I need to talk to a Neopost employee who knows the working of the DS-86 in and out. This will probably take more than just a visit to Neopost. Although I'm not really liking the idea, I need to spend some time at Neopost Drachten. Can you arrange a person with which I can discuss my model during the length of my Masters Project? Our idea is to come to Drachten 3 days every two or three weeks till January.
  • A(D): It is better to like the idea, otherwise you would be an unpleasant exception. Describe what you need to Joost, he will give you the names.

Questions: 08-09-08

  • Q(J): In a week or 2 I'll be ready to come to Drachten to tell about my work so far and ask questions. I would like to pick a date. Any date except a Thursday should be fine with me.
  • A(DJ): The candidate dates will be known on 16th September.

Questions 29-08-08

  • Q(J) Please check the goals (preliminary) page again and take a look at the smaller Neopost goals at the bottom.
  • A(DJ) I intend to do have that done before Thursday 4. September 2008.
  • Q(J) How exactly do you want me to model the DS-86? Maybe I'm thinking of other ways of communication between the Feeder and the Inserter than is implemented in the DS-86. In other words: how true to the working of the DS-86 should I stick?
  • A(DJ) This will be a part of Neopost expectations description of the project, and of our discussion on Friday 5. September 2008 in Eindhoven; but for the beginning: you can start with specifying interfaces in the way YOU THINK THEY MIGHT WORK.
  • R(J) Maybe I'll put some more questions up after my meeting with Bert this friday afternoon. Normally this happens on Thursday, but I was busy getting my Motorcycle license (which I did :)), so we had to postpone the meeting.
  • A(DJ) Congrats! Go on with posing questions, as detailed as possible!

Questions 19-08-08

  • Q(J) Please check the goals (preliminary) page for the goals I have set so far. Please tell me if your (Neopost) goals are included. This is for the full project assignment we want to have before the TWINS meet of 5 September.
  • A(DJ) The peliminary goals description as presented at this moment is an excellent starting point, and it lists all the relevant aspects; the quality is satisfactory, the quantitiy as well. However, I want to make some 5-8 smaller and simpler point, for an easy measuring of the progress and the final result; this practice became standard for Neopost internship projects. But you can stick to by you proposed list for the TU/e goals. I will propose my bullets before 5th September… and beware the ambition “till all people involved are satisfied” :-)
  • Q(J) Is there still need for an update of the NTBV ↔ SE Tu/e document? If so I'm the person to give the input from the SE section.
  • A(DJ) Yes please, fill the parts of that document with with everything known by you at this moment.
  • Q(J) In your documents and some of you answers on this page the terms 'syntax' and 'semantics' are used. I basically know what the meaning of the words are, but in what meaning are you using it here? I know how a programming language is highly dependant on good syntax and semantics, but I don't think that's the discipline I should be thinking in. Could you help me out here?
  • A(DJ) I do not stick here to any particular definition here (but we can and perhaps should select something from Wikipedia for instance), but to the “common sense” that 'semantics' implies “meaning and properties of that meaning” and 'syntax' the way how to describe that meaning. This is the same notions like in the programming lang's, why not. In the NeoMultiModel you have those drawings as a proposed syntax, but what they actually should mean (drawings) - that is one of the subjects of this project.

Questions 15-07-08

  • Q (Jelte): In ICSSEA section 3 you sum up 5 points as RE objectives. In point 4 you mention: “to ensure completeness and consistency of these requirements,…”. However requirements aren't mentioned in the above 3 points, don't you mean “properties” there?
  • A (Dusko): Yes; here I forgot a little bit that there are also people on the Earth which are not familiar with the RE. Actually, a product “properties” are a subset of a product “requirements”; so, when I wrote “these requirements”, I meant “the properties as a part of the requirements”.
  • Q (Jelte): For this project I would like to use bits of the models mentioned in ICSSEA, however these won't be fully usable. How strict would you like me to follow the NeoMultiModel, the workflow in figure 4 and the graphical entities?
  • A (Dusko): Specify “usable”. The NeoMultiModel is, in my opinion, an ideal to strive to - the closer you get to the ideal, the happier the NeoMultiModel stakeholders will get. All solidly justified deviations are equally welcome; time limits are also not forgotten.
  • Q (Jelte): To have enough work during your (Dusko and Bert) vacation I would like to give some points I'll work on during this vacation (Please give your opinion about these relevance of these points at this early state):
    • Describing the black box behaviour of the DS-86.
    • (DJ): relevant. Curious how to do it.
    • Splitting the behaviour in 2 parts (the feeder and the inserter), based on the machine.
    • (DJ): relevant. Especially to see what a behaviour will emerge when this two put together.
    • Defining an interface to have these 2 parts communicate, based on the actual interface.
    • (DJ): relevant - clearly from the NeoMultiModel motivations.
    • Making a chi-model for the two parts and linking them together with the interface.
    • (DJ): relevant - see previous.
  • Q (Jelte): The interface part will be the hardest part to model, can you give me some info on this interface? I read in the service manuals that every sensor and actuator is connected via the interface to the inserter mainboard, so that their states can be checked via the screen. However most of the sensors aren't necessary for the machine to operate properly.
  • A (Dusko): About the interface, my knowledge (including the documentation I shared with you) is the same as yours. What one wants to achieve with interface modelling is also a better understanding of the system. Why would not you apply the philosophy from Figure 6 (ICSSEA), model by Chi only surface (“black box”) behaviour of the two modules so you can see how they are progressing in time by interacting with each other; later add the interface to the user. Pick up a job on DS-86, and look what interaction takes place. How does the folder know when to eject a folded set? Does the folder unit constantly deliver mech. power to the inserter? What interface is necessary for the folder to inform inserter how to open an envelope? Also make assumptions how does this work (taking into account all available information and experience), then come in Drachten to present your insite in the machine logic. If you model is adequate, you will efficiently gather feedback from the experts to further improve your model. Then you can go one stap further (or “lower) in Figure 6 and try to replace the surface behaviour description of “aBigComponent” with some module internal division. Then present your perception of the system; and then you get feedback… This is stepwise refinement - a hot topic already for many years.
  • Q (Jelte): Do you have the graphical entities in a Visio template file, so I can easily make the graphical representations?
  • A (Dusko): I will send you Visio drawings; I have not yet made a templates of the drawings.
  • Q (Jelte): I got a letter from Neopost with a new 'geheimhoudingsverplichting'. In the letter it says that I'm expected at 0900 on the 28th of July, this is probably a misunderstanding or part of the standardletter. But please tell S. Woudman - v.d. Valk or somebody who concerns this that I will not be there on that date.
  • A (Dusko): Don't worry, earlier it has been communicated with the secretary with Mrs v.d. Valk not to expect you on that day.
  • R(emark) (Jelte): Please send me the last draft of [CBD-NTBV] as mentioned in the answers of the last question round.
  • R(eaction) (Dusko): Please check if you have received email “RE: SE Wiki page” from 20080711@1025.
  • R (Jelte): Because this is a somewhat public wiki and I am working on something of which parts are confidential, please let me know if, by accident, I put something up here that's classified. Please edit it out and mail me why.
  • R (Dusko): Since DS-86 is a known product, there is a substantially less concern of the confidentiality breach. We will share the responisiblity here: I will censor the published material, you will indicate any somewhat greater mistake potential.
  • T (ip): Dutch comics right? Check this one: Dirk Jan. Yes I´ve read your home page more or less =)
  • T(hank): I should edit that page to put there now “comics in French and Spanish” - good Anylanguage comics are always very welcome - thanks!!

Questions 11-07-08

  • Q (Jelte): In the document for ICSSEA, section two, you talk about highly abstract, yet domain specific models, based on which four points should be possible. I don't fully understand the use of a model there. Can you give the definition for a model in that context?
  • A (Dusko): Figure 4 in that paper gives the context. The mentioned model is a structural-behavioural description of a system above the step “Monodisciplinary component design (Modelling component in discipline-specific tools)”. The use of the model is in those four points. We search for a syntax and semantics which would allow us to deploy the workflow on Figure 4 in modelling, and allow us using the model for those 4 points in section two.
  • Q (Jelte): In section 4.1 of the NTBV ↔ SE TU/e document the document [CBD-NTBV] is mentioned, to better understand point one I would like to get a copy of that document.
  • A (Dusko): this was a predecessor of the “Requirements for Multidisciplinary…” (ICSSEA) paper. In [CBD-NTBV] there are some “thinking aloud” parts, the ICSSEA paper is a much more consistent piece of text; I consider it the v.1.0 of the [CBD-NTBV], but I will send you the last draft of it (v.0.6).
  • Q (Jelte): The program NeoFlow, and with that the paperflow, seems very important in designing a new machine, should I focus my attention on the paperflow through the modules or on the modules itself? In other words: should I make a process for every mailpiece or for every module?
  • A (Dusko): both approaches have been used for MDD in Neopost. MSc student Ben Sikkens (RuG) has made in UPPAAL a model for mailpiece. He found it useful to interact with models of other parts of the mechanism. In NeoFlow models however the mailpieces are not separate entities, they are part of a module description.
  • Q (Jelte): The DS86 consists of two modules which are essential for the process, the feeder and the inserter. There are however also other modules like the VersaFeeder. In my opinion I should just start with the feeder and the insert, but maybe Neopost wants me to add some of the other modules as well?
  • A (Dusko): It is sufficient to focus now on the two main modules. If your approach is good, however, at any point in time the model should be straight-forwardly extensible. We will surely come to this point in the future.
  • Q (Jelte): Could you put your planned vacation into the google calendar, so I know when I won't get any answers. If you don't want to use Google Calendar, just give the dates on this page.
  • A (Dusko): I added the calendar, and filled some in (your URL for “Import via URL” worked well - thanks!)

Intro Question

  • Q (Jelte): Is this what you expected of the Wiki page?
  • A (Dusko): It's good! This will be the root page, there will come a few other content pages, not too many! Good job, terribly good busy (“erg goed bezig”)!
twins/master_jelte_leijenaar/questions.txt · Last modified: Wednesday, 22 April 2009 : 14:51:32 by djov