-
Notifications
You must be signed in to change notification settings - Fork 102
Background
The links below describes the background to R. How did it start and grow, and how was it transformed from Instat into R-Instat. We also try to answer some questions you may have on R-Instat
Instat before R Background to R-Instat ›› Parent Previous Next
Instat was written by statisticians engaged in teaching, research and consultation, in collaboration with programmers. It began in 1983 as a series of programs for the BBC Microcomputer for teaching statistics, written by Bob Burn. He later worked as a Principal Statistician for the Statistical Services Centre following many years working as a consultant in Mauritius.
Instat was first used on a training course of statistics in agriculture given in Sri Lanka at the end of 1983. It was then further developed and marketed from mid-1985.
In 1987 Instat became available commercially for the PC compatible range of machines. The first shareware version (V6.1) was released in 1994, V6.2 released in September 1996 contained many new general features, while V6.5 had new chapters for the Instat Climatic Guide and new commands to model rainfall data.
Instat was used in the UK and abroad by a wide range of companies, research institutes, schools, colleges, universities, and by private individuals working at home. It has been used in many courses on general statistics and on statistics in agriculture, health and climatology courses, given in Reading, Sri Lanka, Africa, China and elsewhere.
During this period, development of the package continued in collaboration with staff from The University of Reading under the guidance of Roger Stern, who returned to work at Reading, after seven years at ICRISAT in Niger in the 1990s. During Roger's absence, Joan Knock ran the Instat office at Reading.
The first test release of Instat for Windows was in 2000, in time for it to be tested and then used on 10 week training courses in Nairobi in May 2000 and Reading in July 2000. The first public trial version was released in October 2000, with further releases in Summer 2001 and September 2002.
These versions were produced by Roger Stern, Colin Grayer, Ian Dale and Joan Knock, all staff at the Statistical Services Centre, The University of Reading. Funding for the first release under Windows was provided by the UK Met. Office at Bracknell and the Statistical Services Centre. The Statistical Services Centre funded the development of the 2002 release.
Early versions of Instat were totally in BBCBASIC(86), with some sections in Assembler - thanks to Richard Russell (who was primarily responsible for BBCBASIC on the PC). The Windows version added a front end, written in Visual Basic.
Since 2003 successive versions of Instat were in Windows, using versions of BBCBasic for Windows (http://www.bbcbasic.co.uk/index.html ) for the Instat server, with the front end remaining in Visual Basic.
For the past 10 years the main users of Instat have been those who wished to analyse climatic data.
We bow to the inevitable in moving Instat from BBCBasic to R, but those who remember the BBC microcomuter and its BBC Basic language will appreciate the nostalgia that some of us feel, particularly as BBC Basic continues to be available.
R-Instat Copyright (C) 2015-2016 African Maths Initiative (AMI)
R-Instat is a front-end to R. It builds on many of the ideas behind the original Instat, though none of the code from the original software has been transferred. It hardly has a history!
R-Instat was conceived in early 2015, and the development was then made possible through the success of the crowd-sourcing campaign in mid-2015. Development started in the Autumn of 2015, and the first beta relaease of R-Inst5at was on 8 November 2016. Many readers will know of this date as the occasion of the US election competion between Hilary Clinton and Donald Trump. We will remember it as the one year anniversary of the end of our crowd-sourcing campaign and hence as the data of the release of the first beta version of R-Instat.
Applied Statistics at the University of Reading has a long history of innovation in the teaching of statistics. We give some information here and the R-Instat Introductory Guide will also include some of our ideas on the teaching of statistics.
In Africa, and perhaps also elsewhere, Universities and other organisations do not always have an easy-to-use statistics package that can be freely distributed to their students. Lecturers sometimes cite this lack to explain why they are still teaching in traditional ways - ways that over-emphasise formulae, and that students often claim are too theoretical. They also over-emphasise the methods and formulae that were already available in the 1960s with courses largely following the order of mathematical complexity. This is no longer needed, and courses can now concentrate more on ideas and concepts. In R-Instat we have tried to show what this can mean in practice.
In addition, a greater use of realistic sized data sets, could tempt some courses to spend far longer on aspects concerned with organising and describing data (graphs and tables). These are a large part of the practise of statistics. They are also fun and are sometimes largely absent in training courses.
When the Statistical Services Centre were the Biometric Advisors to the DFID Natural resources Division we produced a series of "good-practice" guides. We are grateful for DFID's encouragement to distribute this information freely.
A series of Case Studies, to show the varied roles of a statistics package in teaching was added to Instat as long ago as 1990, when a Teaching Pack was produced.
This menu collects ideas and resources concerning the use of the computer to support the teaching of statistics. Most of the resources are also freely available outside Instat for trainers who wish to make use of them while continuing with a different statistics package.
The motivation and contents of this menu are described in more detail under the Strategy option.
The Case Studies provides a set of examples of ways in which the computer, and a package, such as R-Instat can be used in teaching. These case studies can be for reading, though the descriptions have yet to be updated for the Windows version.
All the datasets are also provided so students or teachers can evaluate the ideas.
Finally we list other resources that may give ideas for trainers and trainees in relation to their statistics teaching. These are partly from Reading, but mainly from other sites that we have found useful.
Since 1986 special modules were included for the analysis of climatic data. These commands were written by Roger Stern, Hugh Hack and Joan Knock. Further advice and help was provided by James Morison and Mike Dennett, from the University of Reading.
Instat was first used outside Reading on a World Meteorological Organisation (WMO) course in CIMH (Caribbean Institute for Meteorology and Hydrology) Barbados in October 1988. The Climatic Version of Instat and the Climatic Guide were then produced in 1989 and Instat became used by Meteorological Services in many countries, particularly for agroclimatic analyses.
The Climatic Guide was translated into French and Spanish in the early 1990sby staff from the Algerian and Costa Rican Met Services. The climatic facilities of Instat have been used on courses at Reading and on regional courses organised by WMO, Geneva in, for example Burundi, Panama, Philippines, Syria, Algeria, Niger and Brazil. Since 2000 they have been used on regional courses in Kenya and Niger. The software has also been used routinely by the climatological sections in many Met services for the analyses of their historical climatic data
We also acknowledge support from the UK Met. Office over many years. This included the resources to upgrade the climatic facilities in Instat to Windows.
Since 2012 staff at the University of Reading has developed the PICSA approach (Participatory Integrated Climate Services for Agriculture) . This makes extensive use of the historical climatic data and Instat has been used routinely by National Met Services for this work. It will be superceded by R-Instat from early 2017.