Martin and Christoph had the idea for a dedicated face-to-face version of the MarlinTPC meeting concerning the /reconstruction for the large prototype/. 
Very quickly it became clear that this meant to be a sort of get-together what actually needs to be done in terms of software and program flow.
Depending on the response we suggest a meeting of three to five days in Hamburg/DESY, proposal for the time is either the week starting 14. September or the 21. September.

A couple of topics quickly appeared:
* A hands-on introduction into MarlinTPC, what steps are actually needed to get from the detector data (e.g. ADC counts per channel) to higher steps in the reconstruction from Pulses to Hits to Tracks.
* In which shape is the reconstruction chain within MarlinTPC, e.g. what processors are there, to which extend are they functional and what is still missing.
* How can the comparative analysis of collaboration data, e.g. testbeam data, be organised.
* The setup of information systems, e.g. databases concerning the data like: file location, runs, settings (gain, gas, etc.), geometries, slow control data, etc.
* /HowTo/ use the grid for data analysis.
* Open issues/future direction for MarlinTPC and its use within the collaboration.

Further topics are of course possible and are explicitly requested for external proposal.

Feedback is needed:
- for the date (are there any conflicts for anyone willing to attend)
- on the topics