John H shared a video he had made of a student presentation on prototyping in 1987; students had been divided into groups of four to research a topic and his group had decided to present their results by way of a series of sketches. At the time development mostly involved COBOL and programming only started after the requirements had been fully specified which normally meant that, by the time the program was delivered, things had moved on and the program no longer met the needs of the organisation. The proposed solution was prototyping of a model of the program to get user feedback before embarking on the programming or building the entire application by prototyping through a series of iterations in much the same way as free and open source software is now developed.
Brian demonstrated in this presentation how he set up a Raspberry Pi 2 as a router essentially by setting up a static IP during configuration and then handing over to a DNS server.
John W asked the best way to link together a number of computers and NFS or Samba were suggested.
John H talked about the background to and the work of Claude Shannon, the centenary of whose birth fell on 30 April 2016.
David S had hoped to be able to demonstrate BASH for Windows though he had found that he had to sign away all his rights to register on the Windows Insider Program and, when he had done that, found that the relevant option had not being installed on his tablet. So all he could do was point to the BASH on Ubuntu on Windows site.
John H picked up on a discussion at the previous meeting to give a presentation on dBASE II. He had never upgraded to dBASE III because it was not backwards compatible with dBASE II (other programs of that era like WordStar and Supercalc had maintained backwards compatibility; so it was possible to use them on both CP/M and DOS machines) and because dBASE II had an operator similar to
LIKE "%<substring>%" in SQL which had not been implemented in dBASE III. As he had made extensive use of this operator in his programs, an upgrade to dBASE III would have involved an extensive rewrite of all his programs.
Alice started us off with Optimising Impala Queries, or a ‘Distributed Lego Community’, a demonstration of the principles behind Parquet, a columnar storage format, and Impala, an analytic database, for the Hadoop ecosystem. Columnar storage formats overcome the burden of reading every row of a table based database such as SQL.
With no presentations prepared, we talked about this and that, from Elixir, a concurrent programming language that runs on Erlang, to openSPARC.
Brian introduced us to the Pine 64, an expandable single board computer starting at $15 for 512MB. Though a 2GB version was advertised, it appeared that only the 512MB and 1GB versions are currently available.
Stephane then recommended the Charbax videos and in particular the interview with Bernhard Rosenkränzer on the Android team at Linaro and Rob Clark of Red Hat who works on the open source GPU driver called Freedreno for Qualcomm’s ARM processors’ Adreno GPU. He noted that ARM GPUs are all bound to specific implementations of the GPU which makes producing common code very difficult.
John H began with a presentation on the background to the recent release of openSUSE LEAP 42.1.
Alice then demonstrated using the Realtime Trains API to download and analyse information about train movements on Train workings; the source code is on GitHub.