For 35 years, the field of machine translation (MT) has had its own journal, with the first issue appearing in January 1986, with three original papers, two book reviews (in one contribution), a paper summarising three recent conferences, and a ‘brief communication’ noting how the world was ripe for MT even back then! The original papers focused on post-editing of MT output, sublanguage MT, and interactive MT, all hot topics still today! Plus ça change!

It wasn’t called Machine Translation then; it was called Computers and Translation, hence COAT, which is why even now if you have ever submitted to the journal, the number assigned is always preceded by “COAT”!

In all that time, there have only ever been three editors: first, Sergei Nirenburg, then Harold Somers, and finally myself. I, and we, are very grateful to them, and to Springer, for setting the journal up in the first place, and with such firm foundations that it allowed me to be the bearer of the baton for the past 14 years. In all relay races, there are 4 runners, not three, so I am pleased to be able to pass on the baton to Maja Popović as she takes care of the field as we enter this new journey with Language Resources and Evaluation (LREV).

It may appear that I am also the captain going down with the ship, and that the apparent demise of the journal is a loss to the field, but we both firmly believe that researchers, users, developers, vendors, and indeed anyone with an interest in the topic shouldn't feel that way. Far from being a disadvantage, aligning ourselves with LREV allows MT articles to appear alongside a wide spectrum of other papers documenting practically all applications involving human language, and of course, nowadays the underlying technology used in those applications is very similar, so there is a lot to be learned from disciplines which traditionally have been separate from MT.

Accordingly, MT practitioners will become more connected with other related fields, seeing cutting-edge research in other areas of computational linguistic/natural language processing of potential benefit to MT. They will also be able to submit their work related to cross-lingual NLP applications such as sentiment analysis and hate speech detection to name but two. The change will also benefit the journal: while submissions about MT have been welcomed all the time as a sub-field of NLP, more MT practitioners will be motivated to submit their high quality work, not related strictly to MT but also connected with other NLP sub-fields. In this way, the journal will broaden its scope and audience. Of course, this is a two-way street; many of these other fields have benefited from improvements in MT, as a particular instantiation of what has come to be called sequence-to-sequence learning in this age of AI.

This mutual appreciation of what is happening in a range of related disciplines will benefit from papers on MT appearing much more regularly than heretofore in LREV, a high-impact journal, note. Whether we like it or not, metrics such as impact factor and FWCI are being increasingly seen as important measures of research activity, especially by funding agencies; despite huge efforts over the past number of years, the MT journal was unable to obtain an impact factor, so unfortunately some researchers simply could not submit their work to our journal. That is no longer a problem.

My final task as editor of the journal of Machine Translation, then, is to wish Maja all the very best in her new role as associate editor for MT for LREV, and to encourage you all to support this move by contributing to LREV by writing scientific papers and book reviews, reviewing papers, and submitting ideas for special issues in the broad area of MT. Our field is extremely well placed to succeed, and needs all our support if it is to continue to thrive.