http://journals.jevon.org/users/jevon-phd/ Jevon Wright's Journal en-us Sat, 06 Sep 2008 07:25:23 +0800 http://blogs.law.harvard.edu/tech/rss journals.jevon.org widget jevon@jevon.org support@jevon.org http://journals.jevon.org/users/jevon-phd/entry/16601 Well I am in Auckland, along with Aaron Steele. This afternoon I gave my presentation but accidentally gave the wrong version!! An important lesson in never carrying more than one copy of a given presentation around. The presentation itself went pretty well... everybody seemed to be interested (at least most eyes were not looking sleepy) and all my feedback has been really positive. I have a couple of hours free in a really insecure Auckland Internet cafe (with everything in Korean - thank goodness for VPN + remote desktop + ProcEx.exe) so I will soon put the presentation online. I also wish I had the time right now to stop those damn spammers >_> Mon, 01 Sep 2008 18:06:38 +0800 http://journals.jevon.org/users/jevon-phd/?showid=16601#16601 http://journals.jevon.org/users/jevon-phd/entry/16601 http://journals.jevon.org/users/jevon-phd/entry/16416 After such a long and challenging process, my project website openiaml.org is now up and running. It took about eight weeks to get a virtual Linux server installed (blame it on the community process -- it took ages for them to release!) and another two weeks to convince Massey University to let me open up some pretty key ports (80, 8080, 443). It's already got heaps of content on it, including a live demonstration installation of my Ticket 2.0 implementation, written in Symfony. More information is on the website including the technical details needed to implement it directly, but I would appreciate others to play around on the website and let me know of any bugs or suggestions! Mon, 25 Aug 2008 10:02:42 +0800 http://journals.jevon.org/users/jevon-phd/?showid=16416#16416 http://journals.jevon.org/users/jevon-phd/entry/16416 http://journals.jevon.org/users/jevon-phd/entry/16415 It's been a while since my last entry. I am very busy for the next two weeks - I have a short, ten minute presentation to give in Wellington on Thursday, briefly discussing my postgrad research. It's part of the IEEE so I'd like to get my work out there with other students. Next week is my WISE 08 presentation. I'm looking forward to presenting it -- somewhat -- because there is some interesting content in there, such as the preliminary metrics from my Ticket 2.0 implementation, that may be quite interesting to discuss with fellow web engineers. I'm also looking forward to meeting with other researchers in my field. I also have my first DOI! :D Mon, 25 Aug 2008 09:58:27 +0800 http://journals.jevon.org/users/jevon-phd/?showid=16415#16415 http://journals.jevon.org/users/jevon-phd/entry/16415 http://journals.jevon.org/users/jevon-phd/entry/16191 After finally replacing our old DSL modem with one that can actually support broadband without crying, I've had the opportunity to install and play around with FreeNet, the alternative content publishing network. It's been giving me some excellent ideas, with regards to my PhD topic. In particular, I am really keen to see if FreeNet will in the future support some form of dynamic content (currently only static content is available). Once the FreeNet can support dynamic content, then it is likely I will be able to use my internet application modelling language to automatically generate FreeNet applications! Mon, 21 Jul 2008 12:16:43 +0800 http://journals.jevon.org/users/jevon-phd/?showid=16191#16191 http://journals.jevon.org/users/jevon-phd/entry/16191 http://journals.jevon.org/users/jevon-phd/entry/16183 Is it possible to model a system's user interface in UML? This seems to be a difficult topic to address. There is no obvious "user interface model" in UML. I think this is because the role of UML is higher level than simply modelling an implementation -- models in UML are designed to describe what you want modelled, not how you want it modelled. That is, UML models describe the interaction and components of a user interface, but not where they are situated or how they look. I think the best way to model a user interface in UML is to model the interactions you would do with the interface, using an activity diagram: You can then combine this activity diagram into another activity diagram with swimlanes, or an interaction diagram with boundary classes, in order to understand where the user interface would actually lie, and the responsibilities of each system. It's not really possible to do a model like this in pure UML: Another option is to specify the component composition of the user interface... Wed, 16 Jul 2008 12:33:49 +0800 http://journals.jevon.org/users/jevon-phd/?showid=16183#16183 http://journals.jevon.org/users/jevon-phd/entry/16183 http://journals.jevon.org/users/jevon-phd/entry/16178 Is it possible to model conventional atomic transactions in UML? Many existing approaches deal with the modelling of business transactions, but usually in custom XML languages, such as WS-TXM, WS-BusinessActivity (Microsoft) and BTP (Oasis). These approaches are more like keeping track of the state of a system though. The best approach that I can find so far in the UML 2.1.2 superstructure is to use interruptable activity regions (pg 379): Tue, 15 Jul 2008 07:49:31 +0800 http://journals.jevon.org/users/jevon-phd/?showid=16178#16178 http://journals.jevon.org/users/jevon-phd/entry/16178 http://journals.jevon.org/users/jevon-phd/entry/16140 One point I haven't been able to properly answer yet is whether to use UML or not for my modelling language. I have two clear options: either extend UML with a UML profile; or create a new DSL (based on MOF or Ecore). It is fairly clear that UML can definitively describe any web application , but UML does not precisely map to executable applications. While UML may be able to describe an entire web application, it is still up to the target platform (usually a user, or a CASE tool) to translate these requirements into a working application. Theoretically, a DSL and a UML profile can be identical to each other. Theoretically, any DSL can be described in a UML profile. A UML profile encourages using UML syntax, and everything is decomposable to core UML concepts, but this comes with a massive amount of additional complexity. Both a DSL and a UML profile would require a custom CASE tool. The only advantage here is that extending a UML tool would provide a lot of built-in functionality, but would also induce a l... Fri, 27 Jun 2008 09:32:41 +0800 http://journals.jevon.org/users/jevon-phd/?showid=16140#16140 http://journals.jevon.org/users/jevon-phd/entry/16140 http://journals.jevon.org/users/jevon-phd/entry/16122 For the last hour or two I have been investigating Open Source licences for my project. In the end I decided my favourite licence option is the Eclipse Public Licence (EPL) (which would integrate nicely within Eclipse itself). Google Code only has a small range of licences it can offer to hosted projects; but unfortunately, the EPL is not one of them, even though it's been requested before. So I've had to move IAML from Google Project Hosting into a new Sourceforge project. I will miss the simple Google project interface, but this is a necessity until Google Code supports the EPL. Fri, 20 Jun 2008 12:04:58 +0800 http://journals.jevon.org/users/jevon-phd/?showid=16122#16122 http://journals.jevon.org/users/jevon-phd/entry/16122 http://journals.jevon.org/users/jevon-phd/entry/16120 This week I have been playing around with GMF and seeing if I catch up on the learning mountain, so that when implementation time comes around, I won't be overwhelmed. It's taken me a few days but I've managed to create a modelling tool that, given a model like so: Can be right clicked and selected to Expand: And then right clicked again and selected to Shrink back: I am curious to see if I can automate this via JUnit testing -- this would allow me to programatically make sure that my implementation won't create errors or an improper model. Also, I installed Firefox 3 yesterday. It's pretty decent. Not as pretty but it's a little more sensible. Also, you can click on <input type="file"> fields to select a file, instead of having to click the button next to it! Thu, 19 Jun 2008 10:03:50 +0800 http://journals.jevon.org/users/jevon-phd/?showid=16120#16120 http://journals.jevon.org/users/jevon-phd/entry/16120 http://journals.jevon.org/users/jevon-phd/entry/16103 KM3 is quite a silly idea. They describe it as a "language-independent meta-model for writing meta meta-models." Um, hello? Considering "meta-model" is practically synonymous for "language", a better description would be: "a language for writing meta-meta models." KM3 is itself a language. None of this language-independent language silliness! Fri, 13 Jun 2008 11:57:30 +0800 http://journals.jevon.org/users/jevon-phd/?showid=16103#16103 http://journals.jevon.org/users/jevon-phd/entry/16103 http://journals.jevon.org/users/jevon-phd/entry/16095 What ever happened to doing entries regularly?? lol After finally getting my marking from this semester out of the way (looking forward to getting paid!!) I've spent this week actually doing real research work. And loving it! This week I have been looking at existing modelling languages for everything from software design and architecture, to events, rules and user interfaces, and trying to decide if I can extend on one of these or if I will have to develop my own. So far my favoured approach is a combination of various languages. I have a meeting tomorrow with my supervisor where we will discuss if this approach is desirable. I've registered a new website for my project -- http://openiaml.org -- which hopefully will have an actual website behind it in the next two weeks, once I get the server set up through Massey. One of my goals of my PhD is to actually use my research to redevelop the Journals. Not only will this prove as a valid research outcome, but it will also finally let me update the Journals fo... Tue, 10 Jun 2008 16:44:41 +0800 http://journals.jevon.org/users/jevon-phd/?showid=16095#16095 http://journals.jevon.org/users/jevon-phd/entry/16095 http://journals.jevon.org/users/jevon-phd/entry/16084 While having a break from marking assignments, I came across this fantastic new Google tool... Google on John Key Very interesting... Thu, 05 Jun 2008 11:57:53 +0800 http://journals.jevon.org/users/jevon-phd/?showid=16084#16084 http://journals.jevon.org/users/jevon-phd/entry/16084 http://journals.jevon.org/users/jevon-phd/entry/16046 Good news! After probably around 3 or 4 weeks of implementation in Symfony, I have succesfully implemented all of the requirements of my benchmarking application (Ticket 2.0) using PHP. Some of the techniques I utilised are available through my del.icio.us feed for ticketsf. My next task is to finish the final version of my paper for WISE 08, as one or two of the requirements have changed, and to also get the application online the Web somewhere. But before I do that I need to come up with a name for my project. I'm terrible at coming up with names so any sources of inspiration would be appreciated :D. Tue, 20 May 2008 12:54:25 +0800 http://journals.jevon.org/users/jevon-phd/?showid=16046#16046 http://journals.jevon.org/users/jevon-phd/entry/16046 http://journals.jevon.org/users/jevon-phd/entry/16002 I've just found the first sexy cousin primes! (7, 11), (13, 17) (7, 13), (11, 17) The first are cousin primes, and the second are sexy primes. Surprisingly there has been no work on this field in the past. Wed, 14 May 2008 10:12:14 +0800 http://journals.jevon.org/users/jevon-phd/?showid=16002#16002 http://journals.jevon.org/users/jevon-phd/entry/16002 http://journals.jevon.org/users/jevon-phd/entry/15965 I've got some really good news. My research paper that I quickly developed and submitted for WISE 08 in Auckland (Sept 1-4) has been accepted! I'm really delighted that it's been accepted, especially since it is an international conference (France, China, New York), it's in a key conference in my field, and the acceptance rate was fairly low (28%)!! My presentation at Vic went OK. I started off fairly badly - serves me right for adding a slide the day before. But I had some great discussions afterwards. The consensus was that my PhD may be a little too broad in scope, so over the next week I will try and work out what key areas I want to focus on. Tue, 29 Apr 2008 20:38:41 +0800 http://journals.jevon.org/users/jevon-phd/?showid=15965#15965 http://journals.jevon.org/users/jevon-phd/entry/15965 http://journals.jevon.org/users/jevon-phd/entry/15962 As mentioned about a month ago, tomorrow I am giving a seminar at Vic about my work. It's been pretty rushed since the last two weeks I spent down south at the NZCSRSC 08 student conference. I am looking forward to next week when I may actually resume my research! (At 45 minutes, it's also the longest presentation I have given; my confirmation seminar ended up being around 30 minutes.) Mon, 28 Apr 2008 11:47:57 +0800 http://journals.jevon.org/users/jevon-phd/?showid=15962#15962 http://journals.jevon.org/users/jevon-phd/entry/15962 http://journals.jevon.org/users/jevon-phd/entry/15920 Over the last week I have been developing and implementing a benchmark application for interactive web applications. The benchmark combines social networking features with e-commerce functionality, and aims to provide a standard Web 2.0-like example application. Currently I am implementing the initial application in the Symfony framework for PHP for a few reasons: The implementation will be a proof-of-concept, showing what a perfectly reasonable web application is, and that it works as desired.Any differences between the implementation of Symfony and existing modelling languages (WebML, UWE) can be clearly labelled as missing functionality and expressibility.It will allow me to define and prototype some of the weirder parts of the application, such as credit-card billing (through DPS), sending text messages (through e-mail), defining APIs (ICal), visual effects and design (script.aculo.us), application structure, and so on.I chose Symfony in particular because is the development framework I am most experience... Tue, 08 Apr 2008 09:01:03 +0800 http://journals.jevon.org/users/jevon-phd/?showid=15920#15920 http://journals.jevon.org/users/jevon-phd/entry/15920 http://journals.jevon.org/users/jevon-phd/entry/15897 In a few weeks time I will be giving a seminar at Victoria University in Wellington on my Ph.D. topic and progress, you can find out more information on the Vic website. Thu, 03 Apr 2008 11:10:50 +0800 http://journals.jevon.org/users/jevon-phd/?showid=15897#15897 http://journals.jevon.org/users/jevon-phd/entry/15897 http://journals.jevon.org/users/jevon-phd/entry/15851 I'm now successfully confirmed as a full Ph.D. student :D My presentation was fairly concise and I got a bit of good feedback at the end. Now I have to quickly finish a paper for WISE 08, which decided to move the submission deadline forwards three weeks :roll:. Fri, 14 Mar 2008 12:01:34 +0800 http://journals.jevon.org/users/jevon-phd/?showid=15851#15851 http://journals.jevon.org/users/jevon-phd/entry/15851 http://journals.jevon.org/users/jevon-phd/entry/15818 Both my report and presentation for my confirmation seminar is now completed, and I'm eagerly awaiting the presentation just so I can get it all out of the way and get back to research! Thu, 06 Mar 2008 13:48:08 +0800 http://journals.jevon.org/users/jevon-phd/?showid=15818#15818 http://journals.jevon.org/users/jevon-phd/entry/15818