Subsidy Request Lire-2.0
tools for computer/network log file analysis
Summary
Stichting LogReport Foundation is requesting a grant of €38.000 from Stichting NLnet for completion of the development of Lire 2.0. Most of this development work is aimed at improving the Lire framework, making it easier for users to configure and use the software in a way which suits them best.
History of LogReport
Stichting LogReport Foundation was established in August 2000 with the help of Stichting NLnet, and started serious development of the Lire software in October 2000. From October 2000 until July 2003 approximately 30 manmonths of paid effort have been spent on development of the software and documentation, setup of website and online responder service, and promotion at various conferences and on-line magazines. All of this was paid for by subsidies received from NLnet for a total of €253.500. In addition a significant number of volunteers have contributed many small chunks of code for integration into the Lire system.
In 2002 some initial attempts were made to spin off a commercial entity "logreport.com" which would sell LogReport consultancy and develop "specials", while continuing to contribute to the GPL'ed source base. But this did not seem to lead to viable results.
The foundation thus has concentrated on using its remaining resources to bring the Lire software to a state where it can live on from mainly volunteer contributions. The internals have been reworked for more flexibility, and to make it easier for less experienced developers to contribute code like new converters and analyzers.
Current status
Lire 1.3 has been released in May 2003 and represents the most current version of the software (a small number of corrections have been committed to the CVS tree since). Judging from the number of downloads, the software is reasonably popular, in a wide variety of environments. This is also confirmed by the subscriptions to and the traffic on the various LogReport mailing lists.
The plans from October/November 2002 to complete a 2.0 release in the same time frame proved to be too optimistic. LogReport's remaining funds are close to zero, and an effort has been made to reduce the recurring costs of hosting the web server and log responder by moving it to to NLnet Labs (thanks for their cooperation!). No staff is employed anymore at this moment.
Plan for 2.0
It is clear that the silver bullet for transforming Lire and LogReport into a community-supported project thriving on volunteer code contributions and release management has not been found yet, and probably will never be found since it probably does not exist ... However, we think we can improve LogReport's chances for continued impact on the world of Open Source generic logfile analysis and reporting tools by embarking on a revised 2.0 development plan (attached below). The proposed work is concentrated on the internals of Lire, a relatively complex area for which outside volunteer contributions are hardly to be expected (and have not been received in the past years). A very important byproduct of this work will be a considerable performance improvement, something which is high on the wishlist of high-volume Lire users.
If funds are made available for implementing the 2.0 development work, Stichting LogReport Foundation intends to award a fixed-price contract for this for a total of € 38.400 to Francis Lacoste. Francis Lacoste, based in Montreal, has been on the Lire development team since June 2001, and is the chief architect of the current product. He will employ a second developer, Wolfgang Sourdeau, also based in Montreal, to speed up the work, and possibly improved quality as a result of "pair-programming". Completion of 2.0 is expected well before the end of 2003.
The contract costs have been established like this:
- basic estimate 74 mandays;
- 20% extra for estimation errors, pair-programming etc.;
- 10 days extra for release engineering;
The resulting total of 99 mandays has been rounded up to 100 mandays, or 800 hours, which at a rate of € 48 per hour, result in a total cost of € 38.400. All development items (labelled 1 till 8 in the development plan) are to be considered mandatory; the fixed-price risk is considered to be covered adequately by the estimation adjustments decribed above.
Future perspectives
Several external parties are using the Lire code, possibly enhanced with local modifications, for serious production. These are prime candidates to contribute to sustained development, but tend to be limited to the plug-ins rather than the actual framework. With the revised 2.0 Lire framework it will be easier to sustain this kind of external contribution. Also, the framework will make it easier and more attractive for consultants to apply Lire in customer-specific situations, another impetus for Lire's continued life.
It is expected that one of the members of the Lire development team will continue to function as a release coordinator / patch integrator on a voluntary basis after the 2.0 release. Francis Lacoste has volunteered for this position.
Appendix:
Development plan by Francis Lacoste.