Hostname: page-component-7c8c6479df-94d59 Total loading time: 0 Render date: 2024-03-30T05:26:26.803Z Has data issue: false hasContentIssue false

Editorial note

Published online by Cambridge University Press:  19 December 2018

Rights & Permissions [Opens in a new window]

Extract

Natural Language Engineering really came about from a meeting between Roberto Garigliano (then of Durham University) and myself in his office in late 1992 or early 1993. I had returned to academia the previous year after a spell doing a variety of jobs in industry, and had become aware of Roberto and the Natural Language Group at Durham (just about 15 miles from the University of Sunderland where I was working). Roberto and I discussed several possible avenues of cooperation, including sponsorship by Durham of students on existing Sunderland masters degrees, a joint Durham/Sunderland specialist Masters in Language Engineering (which came to nothing) and a new journal focused on practical, engineering work in the language domain. Incidentally, one of the sponsored master’s students was Siobhan Devlin, now Head of Computing at Sunderland.

Type
Editorial Note
Copyright
Copyright © Cambridge University Press 2018 

Natural Language Engineering really came about from a meeting between Roberto Garigliano (then of Durham University) and myself in his office in late 1992 or early 1993. I had returned to academia the previous year after a spell doing a variety of jobs in industry, and had become aware of Roberto and the Natural Language Group at Durham (just about 15 miles from the University of Sunderland where I was working). Roberto and I discussed several possible avenues of cooperation, including sponsorship by Durham of students on existing Sunderland masters degrees, a joint Durham/Sunderland specialist Masters in Language Engineering (which came to nothing) and a new journal focused on practical, engineering work in the language domain. Incidentally, one of the sponsored master’s students was Siobhan Devlin, now Head of Computing at Sunderland.

At the time (early 1990’s) Roberto felt, with some justification, there was a shortage of journal spaces for language-based work. The only journal with any reputation in the field as a whole was computational linguistics, but at the time that journal had become overly competitive, had long delays before publication, and seemed to have adopted a possibly unofficial policy of prioritising papers which were underpinned by linguistic theory of one sort or another – what Henry Thompson had called some years previously Computation in Service to Linguistics (Thompson, Reference Thompson, Jones and Wilks1983). This of course made it difficult to get into print engineering papers, focused on issues like scalability, software engineering and maintainability, system integration, operational robustness and so on.

One alternative was conference papers, but they generally did not allow sufficient space to put forward the real detail of overall systems as opposed to tweaks within existing frameworks. The other alternatives were Machine Translation, which was too specialised for much of the field; Information Retrieval journals which often set unreasonable evaluation standards for language work at that time; and finally, generalist computing journals, where sympathetic and well-informed editors and reviewers were hard to find. Hence, our belief there was an opportunity for Natural Language Engineering.

Another reason Roberto and I were convinced our efforts were timely was techno-political. A few years previously, the European Commission had produced a document in which ‘Linguistic Engineering’ was defined:

Linguistic Engineering (LE) is an engineering endeavour, which is to combine scientific and technological knowledge in a number of relevant domains (descriptive and computational linguistics, lexicology and terminology, formal languages, computer science, software engineering techniques, etc.). LE can be seen as a rather pragmatic approach to computerised language processing given the current inadequacies of theoretical CL. (European Commission, “Linguistic Research and Engineering in the Framework Programme (1991) 1990–1994”, p7).

Although the terminology was trivially different, it was clear others were thinking along the same lines, in terms of the importance of developing an engineering discipline for processing natural language.

We wrote what I thought was a well thought through proposal. We touted it around various publishers, initially obtaining very little interest. In retrospect, our document had many shortcomings: it was a bit like a very good undergraduate essay being reviewed for a major conference. We learned from the feedback, and eventually got a bite from CUP oddly from Humanities and Language side of the house rather than the STEM side where we would feel more naturally at home.

The initial feedback from CUP was that they might go forward if we could fix various problems with the current proposal: get someone with more standing the field to join the editors, strengthen the editorial board, and so on. I approached my friend and former colleague Bran Boguraev (then at Apple), who agreed to join Roberto and me as founding editors. I also notice to my embarrassment the first editorial board includes a disproportionate number of my friends and former colleagues from my period in Cambridge during the 1970s and 1980s (although of course many had moved on by the early 1990s). Doubtless, I called in a few favours to get their support! And the Syndics gave the new journal the go ahead.

Of course, we then had the next problem: how to get sufficient material of the right standard by the now fixed date of December 1994 for a first issue in March 1995. It is no co-incidence the first two articles were from friends and former colleagues of the editors and were submitted in June of 1994.

In the end, we failed: the slated third piece did not make the grade in time, but fortunately something more suitable appeared in August and was through the review process by early March. I bullied a couple of colleagues from Sunderland into producing book reviews and we had 112 pages (including one blank and one advert) shortly before CUP started seriously thinking about canning the project.

And so we went on: gradually, we built up a pipeline of submissions, strengthened and broadened the editorial board, both in terms of subject expertise and in terms of international coverage, and built up a network of paper and book reviewers.

We experimented with more open reviewing – but abandoned that idea when it became apparent many reviewers were unwilling to act if their names were revealed to authors.

An important innovation came in came in 1997 when we published our first Special Issue: a double issue on ‘Knowledge Representation for Natural Language Processing’ edited by Syed S. Ali and Łucja Iwańska. This started a tradition which continues to this day of planning one or two special issues per year on highly specialised, important and emerging topics, with guest editors, fixed submission and publication cycles and so on. These do not always come to fruition, but have always provided (for me as an editor) a stimulating source of insight into the development of the field, new contacts and almost invariably new streams of material and reviewers for the journal regardless of whether they actually get into print as a Special Issue.

The Ali and Iwańska special also sparked a debate amongst the editors about whether we wanted to see the journal become a special issue as opposed to an open submission journal. For a variety of reasons, we decided to remain an open and continuous submission journal but with periodic special issues never exceeding about 50% of the published material in any given year. In order to achieve that, and to select proposed special issues in a reasonably fair and open way, we decided to launch periodic calls for special issue proposals, a tradition which also continues to this day.

Meantime, the field of Natural Language Engineering was growing, subscriptions, then downloads and citations of the journal grew steadily, and the journal went from strength to strength.

In 2000, it became apparent Roberto was no longer in a position to act as Executive Editor and I (somewhat reluctantly it must be said) agreed with Patrick McCartan – then the responsible person at CUP – to become Executive Editor provided the editorial board and the Syndics approved, which of course they did.

I had the opportunity to build on the excellent foundation laid by Roberto. However, it was clear that the expansion, extension and strengthening of the field (whether you call it [Natural] Language Engineering, Computational Linguistics, Natural Language Processing, Speech Processing, Text Analytics or what) together with the growing position of the journal necessitated some changes.

In an attempt to respond to this, from 2002, we introduced Assistant Editors responsible for specialised subfields, initially (Text) Generation and Speech Understanding and Generation. Personally, I am never sure of the extent to which this innovation succeeded. However, I do think it did an effective job of communicating the range of the journal which was always intended to go beyond the rather similar and specific interests of the founding editors. It thereby may have encouraged a wider range and better quality of submissions and helped maintain and improve reviewing standards in parts of the field with which we were less familiar.

My period as Executive Editor included developing the regular cycle of work of an established and vital journal: renewing the editorial board to ensure it maintained a geographical coverage representing the field; newly emerging and active younger members of the community were brought into the Editorial Board; protocols were established to ensure board members retired in an orderly and acceptable manner; ineffective and slow reviewers were not used, new editors recruited and so on.

All this happened during a period when my own interests were moving away from Language Engineering. From around 1998 or 1999, I had decided to focus on Information Retrieval including Image Retrieval. I also was rapidly promoted at Sunderland, first to full professor and within months to Associate Dean for Research in Computing and Engineering. I could not have coped with the workload involved without the able and effective support of my PA at the time: Lesley Jenkins, to whom the journal owes much.

Then in 2007, it became apparent I would be offered the opportunity to leave the University of Sunderland as part of a management reorganisation. Uncertain about my own future, I clearly needed to find a suitable home for the journal. At that time, the management of the journal was still based very much on paper and a couple of entirely local databases, so it we really needed the office to be in the UK.

Fortunately, it was agreed Ruslan Mitkov was a suitable person to take over as Executive Editor, and I was extremely pleased when he agreed to do so.

The journal has gone from strength to strength during his Executive Editorship: expanding from a 96 page standard issue four times per year to 160 pages six times per year while improving time from submission to appearing in print for papers, and not compromising quality. Working with the team at CUP, we have also moved the journal into a twenty-first century operation with much more automated and easier to use submission and review processes.

The field has also been transformed: with language engineering applications entering the mainstream of computing under titles like “Text Analytics”, “Named Entity Recognition” and so on, in practical use (often unrecognised) by millions if not billions of people every day. I do not want to dwell on the development of the field in this editorial: that is the subject of a piece to be contributed by myself and Yorick Wilks (a founder member of the NLE Editorial Board) which will appear in a later issue of this volume. However, I hope that Natural Language Engineering has contributed in some small way to that development.

In that meeting, in Durham, in the early 1990’s, Roberto Garigliano indicated to me that the journal might be a three or four year commitment. Get it into print, get it established and move on. In practice, of course, that is not what happened. Natural Language Engineering has now been part of my life for over 25 years. I feel that is long enough, and it is time for others to take up the burden. I have therefore decided to step down as an editor at the end of this the twenty-fifth anniversary year. I intend to join Roberto Garigliano as a Founding Editor. I have learned much about science, research, academia, funding and above all Natural Language Engineering over the last quarter century of editing the journal. If anyone wants the benefit of those opinions, I will be happy to share them, but if they do not, that is fine too.

I wish the journal every success in the future. In particular, Ruslan and his team have my best wishes and support. Long Live Natural Language Engineering!

References

Thompson, H. 1983. Natural language processing: a critical analysis of the structure of the field, with some implications for parsing. In Jones, K. Sparck and Wilks, Y. (eds.), Automatic Natural Language Parsing, pp. 2231. Chichester, England: Ellis Horwood.Google Scholar