Previous Up Next
Dutch / Nederlands

Dairy, October 2014



Sun Mon Tue Wed Thu Fri Sat
              1   2   3   4
  5   6   7   8   9  10  11
 12  13  14  15  16  17  18
 19  20  21  22  23  24  25
 26  27  28  29  30  31


Saturday, October 4, 2014

The Conspiracy against the Human Race

This morning I finished reading the book The Conspiracy against the Human Race by Thomas Ligotti, who is a contemporary American horror author and reclusive literary cult figure. I start reading this book on August 20, the day, I received the book in the mail. Although this book is presented as non-fiction, while reading, I often wonderd if it is might be just one of his best horror stories. In fact, (if I remember correctly) Ligotti argues that reality is the greatest horror imagiable. The book is inspired by The Last messiah by Peter Wessel Zapffe, a rather obscure, pessimistic Norwegian philosopher, whoes central thought is that conciousness is a mistake of evolution and that therefore it would be best if we as humans would stop from getting children. This all stems from materialism, which states that there is nothing beyond life and that death is simply the end. There are many good reasons that support materalism, but one can never be sure that there is indeed nothing outside the universe we observe. Because the idea that your existendec ends with death seems to make everything utterly meaningless (or MALIGNANTLY USELESS, as Ligotti writes), many people believe that there must be something beyond the observable world. But if it is indeed the case that everything is just meaningless, and that it is also impossible to add some meaning to life, it can als be serve as a form of (stepping stone to) enlightment because it leaves the Now as anything there is. Ligotty does not talk about this. From what he writes (including a quote from part VII of Confessions by Lev Nikolayevich Tolstoy) one can conclude that he views the 'state' of enlightmens as but one of the ways to live in denial of reality, to silence conscious realisation that life ends in death, that all life ends in death, that our whole universe will one day end of become totally 'frozen'.

I found some parts of the book difficult to read. The forword by Ray Brassier seems to be written with the help of a thesaurus to pick the most uncommon words and expressions.

Palindrome date

Today is a palindrome date when written according to the (D)D-(M)M-(YYY)Y format: 4-10-2014. The previous such date was October 3, 2013 and the next will be October 5, 2015.


Sunday, October 5, 2014

Avoiding merging

I already wrote about merging in distributed version control and the complex problem of conflict resolution in my series of postings about concurrency in modeling (as part of software enginering). When users are changing a model in close operation, pessimistic concurrency control it the obvious choice especially if you have many small objects with relatively little information which are easily created, deleted, split, joined and/or rearrange in a hierarchy. Many source code management make use of optimistic locking and have tool support for textual merging edits in source files. But here the number of objects (files) are small and most edit take place inside the objects without changing the hierarchy (directory structure). A disadvantage of pessimistic locking is that it requires access to a lock server. There is for example a problem with working off-line. This can be avoided by securing some locks before going off-line and simply restricting operations that do not match the locks.

In large scale modelling where groups of people are working in seperation on parts of the model, sometimes without knowing each other, and where parts of the model are combined in a different location, a centralized repository is often not the right solution, and pessimistic locking is not possible and merging seems to be required. But this is not the case. It is possible, and maybe even desirable, to avoid merging at all. In large scale modelling changes to parts of the model are not often exchanged but mostly released in a controlled process often including reviews. In these cases the model is divided in parts where mostly one group is working at a certain part of the model at the same time. This means that pessimistic locking can be applied on the level of parts of the model. But there is a hitch, it is very likely that there are many reference between objects in the various parts that might have some consistency requirements. It is important to realize that a relationship between two objects belong to different parts, must (as an object) also belong to some part, either one of the two different parts or to a third part. When a model is divided into parts, one has to decide where objects that reference objects from different parts should be housed. Dividing a model into parts along authorities (persons or organisations) who are responsible for maintaining them, fits very well with the fourth level of modelling of information systems.

When two parts of a model are released separately, it means that reference between objects (with a consistency constraint) cannot occur in both directions. At first this may seem an unwanted restriction, but also from a semantic point of view it makes sense. If two parts depend on each other (and changes to them are not released as a unit) it means that when a new release for the part being referenced is created, the other parts need to be updated and released again to create a consistent version of the whole model (even if no real changes to the objects in the part need to be made). It is possible to define which releases of all the parts form a consistent configuration.

It is very tempting to consider two consecutive releases of a part as a kind of revisions where objects are simply associated by means of an identifier. This line of thinking suggest that merging is simple and that global unique identifier of objects are central. But this is wrong because it is not possible on the basis of some kind of identifier (name, GUID, URI, IRI, and/or others) to determine if two objects in different versions are semantically the same, unless such an association has been formally defined. In effect one need to define a relationship between the objects of two releases to know how the changes should be applied in all other parts that reference this part. Such a releationship could be defined on the basis of a combination of identifiers, but in many cases, when objects are split and joined between releases, it is not possible to define a one-to-one relationship.

When creating a new release for a part for a new release of a part it depends on, it might be needed that for some time, the part contains references to both releases, to support the process where the references are updated one-by-one to the object in the new release. In many cases this could be an automatic process based on the association defined between the two releases, but for example, when an object is split, it might be that a choice need to made about to which of the two (or more) objects the reference need to be made to. This all may seem to be overly complicated, but I think it is best way to deal with the inherent problems of large scale modelling.


Monday, October 13, 2014

Books

At 12:37, I bought the following two books from bookshop Broekhuis:


Saturday, October 18, 2014

Amsterdam

I went to Amsterdam to visit the exhibition 'RGB exit' on the last day, because it will be the last opportunity to see the seventeen works from the serie 'Kleurverhouding' by Peter Struycken together, because sixteen of them have been sold and there is a possible buyer for the last one.

I arrived just after half past ten at the central station. From there I walked to Antiquariaat A. Kok & Zn on Oude Hoogstraat 14-18. I requested to see a book they had put on boekwinkeltjes.nl with the title 'P. Struycken' from 1994. When I saw the book, I immediately told them that it was from 1976.

On my way to bookshop Scheltema, I also had a look in The American Book Center and Athenaeum. At 12:52, I bought the book Susan Sontag: As Consciousness Is Harnessed to Flesh, Journals & Notebooks 1964-1980 edited by David Rieff, ISBN:9780374100766, for € 6,95.

Shortly after I arrived at Galerie Andriesse•Eyck, Peter Struycken invited me for a drink and a delicious piece of chocolate cake at the other side of the canal at Spanjer en van Twist. The weather was extermely warm for the time of the year and we could site outside with no coat. Not after long he was called in, because some guests had arrived. There were quite a number of people visiting the gallery. I explained the purpose of the painting to some of them. I also followed some of the discusions Peter had with the visitors. He is a very kind man with a very broad knowledge, who has really thought deep about some subjects and has some very interesting ideas to share. I always enjoy listening to him and see him interact with other people. In some way, he is thinking like a scientist, and he did write some scientific publications, at least one journal article but also a book. I understand he is planning to write another book for which he has gathered materials in the past years.


Monday, October 27, 2014

Reborn

This morning, I finished reading Susan Sontag: Reborn, Journals & Notebooks 1947-1963, which I started reading last Tuesday. The book is edited by David Rieff, the only son of Susan Sontag. It really made me think about some things in my own life. It is good that it read "Journals & Notebooks" in the title, because the very varied style of the book. At some times, Sontag gives a very detailed (hour to hour) description of the events, while at other places months pass without any note. Het notebooks also contain list of books she want to buy and/or read. During some periods she only used the notebooks to write down the many movies (sometimes up to three a day) she watched.


Thursday, October 30, 2014

The diamond problem

On wikipedia, the diamond problem is described in the contect of multiple inheritance, a feature of object-oriented programming languages. But it also occurs in the case one has versions of document that reference eachother. For example, take the situation where there is document D that is used as a basis for documents B and C, and there is also a document A that references those two documents. Now lets suppose that there is a new version of D, called D'. This possibly means that documents B and C have to be adapted, if it where only to change the references in case the changes in D' have no effect. This results in new versions of documents B and C, called B' and C'. Now as a final step the document A has to be updated, during which all references to B and C are changed to B' and C'. (This assumes that all references have an explicite version indication, and not something like the last available version of B, B', or any later versions.) In this simple case, when there are only four documents, it is not difficult to determine when one can start working on adapting document A, but in case of large scale modelling, there can be hunderds of versions of documents (or models for that matter) with complex dependencies, if one wants to enforce the rule that always but one version of an external document is being(implied) referenced. By this I mean that it would not be allowed to create a version of document A that references B and C', because B depends on D and C' depends on D', meaning that A would contain (implied) references to both D and D'. I am getting the impression that selecting a set of references that have to be updated simulatiously could even be a NP-complete problem. There are also some practical reasons, why one would like to (temporarily) lift this rule, for example when work on B' has finished but work on C' is still in progress and might take some time, and one would already like to start working on incorporating the updates made in B' in document A. This is another reason why when modeling in the large one should allow a model to contain references to different versions of an object in different releases of model that is being referenced as described as a solution for avoiding merging.


This months interesting links


Home | September 2014 | November 2014 | Random memories