News

Today is our projects 2nd Birthday, in the last 2 years of working on Dohickey the programmers have gone from almost no experence with python or hardware information to having a usable program useful in every day hardware enquiry. So congratulations to everyone who has been with us all this time, and thank you to all of those people who have helped us along the way.

The author of Dohickey lives in Boston and had the great fortune to be available for the entire week of UDS this week. during this time we wanted to make sure that hardware issues are further addressed in up and comming Ubuntu releases and where the Dohickey team could work further with the Ubuntu developers and packagers. In this vain we hope to release feature which enable computer readable fields such that any information relating to a device which can effect it's operation is known to the system. There was also much talk about intergrating more of the Dohickey logic into HAL in order to provide the comprehensive hardware information to all of the information tools that use HAL as their base.

In order to fix a number of peoples install problems where they were using dapper or edgy (python2.4) I have released a point two version without the 0.7 features. please install this version instead of 0.6.1 which may give you errors. this deb also includes proper deps checking and I believe it to be a more stable and ubuntu way of creating the deb (thanks to MOTU group)

OK so I've been lazy, not working on Dohickey for the last two weeks at all other than a few bug fixes; I'm preparing new features for the 0.7 release in an effort to stay my uninterest of the project and work on critical features and bugs. So far one of the problems has been pushing the project forwards without much help or even interest from users or other developers. This is trying my patiance but I can only assume their waiting for a full release with fully working server side; or perhaps it's true that users only like to be spoon fed and don't like collabrative projects that ask too much *shrug* I'll let you know in a few weeks how work is going.

The bug fixes available in this release allow communication with dohickey-project.com to sync hardware information in a sane manner; we can now test the new server side code which should be far more stable and includes a search engine; At last we're almost to Beta.

The new release is available for download, we want as much information and bug information as posible for these release because of the large number of improvements and the increase in dependancies.

The next release of 0.6 is due out soon and I've gotten together the inner circle to test all the installation methods and various packages, while the up comming release will be made without much publication, the release of improved server tools for the adminitration of incomming data will prompt a drive to spread the applications install base and the number of hardware profiles and field structures available.

The required rewrite of the hardware decoding and raw data has been complete. we are now able to convert the research data into xci files (xml computer information) files which at first only contain raw information (bios/hal) but once processed and saved also contain processed information for speed. The improvements in typing, class seperation and processing vs raw data have improved the code base considerably.

We gathered together almost 500 computer profiles which we could use to do research on to determin what kind of variance in the Bios and Hal information we could expect. this has lead to a realisation that the local computer discovory code can no longer be tied into the gui and that the confusion between local computer, hardware items and hardware types will cause problems with new designs, multiple items of the same type and importing of information from other computers.

In order to research the variable nature of bios information I created a script which collects many bits of information about a computer and packages it up. I asked the members on ubuntu forums to post me their computers so I could better understand the differences. Someone posted it to Digg and as a result I now have over 300 computer profiles in order to use for dohickey testing. The code will be shaped so debuggers can use text files from other machines where they want to test those machines allowing greater flexibility. Dohickey won't just display your hardware but now also someone elses.

Recently we have hit upon using the DmiDecode application taking elements of data from the bios in order to afix a computers motherbaord, processors, memory and even system information. research continues on the best ways of normalising the sometimes variable data present in the bios and of fallback methods for computers without a bios (i.e. Apple Hardware)

The website has been moved from sourceforge to dreamhost funded by the core developer DoctorMO, this is both to allow users to access information when sourceforge is down and also to enable the perl server side at our new DNS dohickey-project.com; this should now be our perminiant home for all web based activities.

The documentation of both the dohickey client and the dohickey server will now be hosted on a wiki http://dohickey.wikispaces.com/ this will be kept external from the website for now while wiki software is not installed on this server. links from the client will be created to open a browser window and point to the wiki site for help. please do update the wiki with new and improved information when you have it.

Repackaged to install via python setup.py, can now be run from package directory and includes hardware information screen.

Big step forwards in client side code, we now having the rating system, server side intergration and better server control.

First Release of the perl server side code, only available via cvs, but now good enough to use.

Good News! The site is now pretty much completely cross-browser compliant. It's been tested across three major browsers; Firefox, Konqueror, and Opera. The CVS code on the Downloads page is still a little iffy on Konqueror and Opera, but that is due to the way they handle the pre tag, and there's nothing to be done about it. A new design has been implemented behind the scenes, so as you may have noticed, spacing between the elements and size of the elements is now automated; looks much neater. Expect more cosmetic improvements over the next few days (I know that "stats" box is ugly; working on it). We still recommend Firefox for best browsing experience.

Though the site is now publicly open for use, only Firefox is recommended as a browser. We are working toward low resolution compliance, and still working out some issues with Konqueror and Opera. Internet Explorer is not recommended and will likely not work well with the site.

Fixes a number of issues people were having with device nodes and crashing in SuSE

Fixes install and packaging issues (new Makefile) and a fix for 64bit machines

First Alpha Release contains basic hardware detection, information profiling and configuration.