DansGuardian 2.10 - an informal analysis of 3.5 years of developments
Philip admits he is no good at this sort of thing and asked to go with my suggestion of simply "It's out and it's great and woohoo!"
Daniel would like to reminisce, divulge and predict the following: "Philip told me it was time to make DansGuardian stable-tagged and I told him that I probably aught not simply add a news item saying 'Stable, init. Next.'
First I would like to talk about how things have changed. In the beginning, before top posters and HTML emails and colour TVs, I developed DansGuardian, in my own time, as there was clearly a need for such a product for Linux for education. I used it in my day job at an education solution provider called Beebug. DansGuardian lead to me being head-hunted for SmoothWall. I had great fun doing my hobby as my day job at SmoothWall developing DansGuardian technologies and the SmoothGuardian product.
After a while doing my hobby as my day job destroyed my hobby in my spare time. DansGuardian was still getting developed of course, but only during the day. Things changed at SmoothWall and I quickly got promoted to CTO/Technical Director and SmoothWall grew and grew. Very soon I was able to employ a young and very very talented developer called Philip Allison in order to take over development of DansGuardian. This was at the time of DG2.9. I had written an alpha of 2.9 and used autotools. Phil's first job was to take what I had written and remove the horrid libtool and make it work.
Since then my job consists mainly of various types of management at all levels with a reasonable amount of technical consultancy and planning. So I no longer actually develop DansGuardian - I employ the people who develop it - I personally pay for and sysadmin the web server. I am still involved in the long term planning of DansGuardian and for example attend or head meetings about the plans for DGTNG - the working title for the next major version. We actually have quite a large team of people at SmoothWall including Phil and several other developers contributing to the next version.
Now that my day job has not been my hobby for quite some time my programming hobby has returned and you can see some of my work here: wowace.com/projects/raidbuffstatus/
Before I go further I would like to say what a fantastic success DansGuardian has been and how its development model is one to be emulated. DG is developed as Open Source software. By releasing it under the GPL it means the filtering engine at the heart of SmoothWall Guardian products is tested and bug fixed far better than any closed-source product. Web browsing is so important to every-day life that a bug or incompatibility in a content filter could be disastrously time wasting and using this development model reduces this risk. It also means that those people or organization which are willing and technically able to roll their own solution, can do.
With the Open Source model often comes the communities that build up around the projects. DansGuardian has a thriving user community and many people who put in a hell of a lot of their own personal time to help other users and to test and to bug fix and even to provide features sometimes. The DG community contributes to the success of the project. I would like to thank all the people on the mailing list especially for the support they have provided.
But why has it taken so long you might wonder? The truth is version 2.9 has been production quality for a couple of years now, but we wanted to keep adding features without the work of maintaining a seperate fork. We wanted to keep bug fixing as well as developing the same version. This is more fun and half the work. We could not call it stable if we kept adding stuff so the beta label remained. However because of the size of changes in the pipeline now it makes sense to draw a line and finally mark it as stable and start a new branch with big fundamental changes.
With regards to the future, we have some key areas we want to improve on. The process model and how concurrent connections are handled is going to be improved to dramatically increase scalability. The configuration system will be changed to be much more flexible and ACL-like. The content filtering algorithms will be modified to use the latest techniques to increase speed. There will be better caching of results for performance. The filtering will move to more of a categorization engine. There are other very cool things we will be adding but as yet I can't disclose them. That is all I can say for now and I can't say how long it will be. But very soon we will be starting work on DansGuardian 2.11 which will be the new alpha/beta branch.
DG 2.10. Enjoy."
Page last modified: 08 October 2008 12:27:52