Monthly Archives: July 2016

  • Bing: tin-eared translation

    Steve WoodsLanguage, Oddities, Social Media, Tech

    When it comes to machine translation online, Google Translate and Microsoft’s Bing Translator are serious rivals, not only for custom, but also for the awful quality of the translations they provide.

    Social media platform Twitter has – for reasons best known to itself – decided to use Bing to provide translations of tweets in languages other than the user’s mother tongue.

    However, it’s not very good, suffering as it does from an inability to deal with context.

    Take the screenshot below from a tweet posted by your correspondent earlier this afternoon, who clicked on the Bing translation link out of curiosity.

    screenshot showing dreadful Bing translation

    Bing has managed to mangle my tweet, which contains a colloquial French expression (i.e. “du bidon“) into the incomprehensible “your reporting is Tin“, complete with capitalisation that was not in my original text. Bidon can indeed be a tin – or can or container – in French, but it also has the meaning of belly or stomach too. Furthermore, besides being a noun, bidon can also be used an an adjective, in which context it means phony or bogus.

    Wordreference.com has a brief forum thread on the phrase “c’est du bidon“, which passing Bing Translation developers may like to read.

    In the meantime, if any readers out there are contemplating saving money by using online translation tools instead of a human being, you may like to reconsider.

    Readers interested in why I was retweeting something aimed at the Mail Online website may like to read Tim Fenton’s post debunking the Mail’s Bataclan torture rumours.

  • Blunders at the speed of light

    Steve WoodsBristol, Language, Media, Oddities, Tech

    There was good news this week for Bristol businesses with a yearning for high speed internet connectivity.

    The Bristol Post reported on the deployment of ultra-fast 1 Gbps internet in the city.

    While journalists at the Temple Way Ministry of Truth are quite competent at their main task of churnalism, such as copying and pasting the words of wisdom given in press releases by men in suits – as in the article in question – standards slip dramatically and the absence of sub-editors and the associated lack of quality control are patently obvious when Post staff try simplifying complicated technical concepts, as shown by the following sentence.

    sentence reads These glass cables deliver an internet connection at the speed of light which is highly reliable and efficient

    Shall we just examine the above sentence in detail? There’s plenty wrong with it both technically and grammatically, which schoolchildren sitting their SATs examinations at ages 10 or 11 years would be embarrassed to get wrong.

    Firstly, those glass cables. The proper designation is “optical fibre cable“; and as is well known the correct use of terminology is important. An optical fibre cable is a cable containing one or more optical fibres that are used to carry light, whilst an optical fibre itself is a flexible, transparent fibre made by drawing glass (silica) or plastic to a diameter slightly thicker than that of a human hair. So an optical fibre cable can be made of either glass or plastic, i.e. not solely glass.

    Data from an internet connection is transmitted as light down an optical fibre cable. Light travels at the speed of light. However, it is the method for providing the internet connection which is “highly reliable and efficient, not the speed of light. The subordinate clause, i.e. “which is highly reliable and efficient is misplaced and should at any rate have been preceded by a comma.

    Finally, there’s that speed of light; it’s so reliable and efficient that its precise value is 299,792,458 metres per second (approximately 3.00×108 m/s). It is commonly denoted as c, as in Einstein’s famous mass–energy equivalence formula. Furthermore, c is the maximum speed at which all matter – and hence information – in the universe can travel.

    In the slightly better old days when the Post still employed proper sub-editors, any decent holder of that position would have taken that sentence to bits and re-written it roughly as follows:-

    “These fibre optic cables deliver an internet connection reliably and efficiently at the speed of light.”

    Or alternatively:

    “These fibre optic cables deliver a reliable, efficient internet connection at the speed of light.”

    Unfortunately, local newspapers and their online analogues nowadays seem to have forgotten that quality matters and with quality comes a reputation and with the latter, authority.

  • Spelling IS important

    Steve WoodsLanguage, Media, Oddities, Politics, Social Media

    As someone who’s worked with language for the best part of four decades, your correspondent recognises the importance of correct spelling.

    One area where this matters more is people’s names, something which the fourth estate doesn’t always manage correctly; for instance, a couple of years ago in a piece in the Bristol Post in which I was quoted my surname mysteriously changed from Woods to Wood halfway through.

    However, it’s not just journalists who get proper names wrong. Here’s a fine blunder from former Labour leader Ed Miliband on Twitter.

    tweet in which Miliband confuses home secretary with porn star

    One question remains: would an actress/glamour model make a better replacement prime minister than an authoritarian home secretary?

  • Software Heritage announced by Inria

    Steve WoodsOpen Source, Tech

    Software Heritage project logoYesterday Inria, the French National Institute for computer science and applied mathematics, announced (French press release. Ed.) the launch of Software Heritage, an initiative collect, organise, preserve and make easily accessible the source code of all software that is publicly available.

    Sending messages to family and friends, paying bills, purchasing goods, accessing entertainment, interacting with central and local government, finding information, booking travels: nowadays almost every act of our daily life relies on computers and software.

    However, that is just the tip of the iceberg. Software controls the electronic equipment embedded in the machines we use to travel, communicate, trade and exchange. Software lies at the heart of medical equipment and devices; it ensures proper operation of the energy, transport and telecommunication networks; it powers the banks and financial institutions; software is crucial for the working of large public and private organisations of all sizes, be that on mobile devices or in the cloud.

    In summary, software is a key enabler for all aspects of our modern world: our industry, our science, our lifestyle; all of our society depends on software.

    The challenge

    The goal of the Software Heritage project is to build a modern “Library of Alexandria” featuring software, which will form a unique reference database of all source code, a tool for new software projects and a research instrument for computer science.

    Software Heritage is an essential building block for preserving, enhancing and sharing the scientific and technical knowledge that is embedded to an ever-greater extent in software; it also contributes to our ability to access all the information stored in digital form.

    Software Heritage will adopt a distributed infrastructure in order to ensure the long-term availability and reliability of its archive.

    Software Heritage will provide a reference knowledge base for all open source software used in industry, thus enabling better lifecycle management and long term preservation of industrial software. Once live update capabilities are enabled, Software Heritage is bound to become the reference archive for all industrial users, helping software developers of new software projects find, re-use and archive new source code.

    Software Heritage is the foundation on which we can build a unique research instrument for studying all the software source code, enabling significant advances in all domains of computer science and leading to better quality, security and safety in the software on which we depend in our daily lives.

    As of yesterday when the project was announced, Software Heritage had already collected more than 20 million software projects, archiving more than 2.5 billion unique source files, along with all their development history. Software Heritage therefore represents the richest collection of source code on the planet.

    Antoine Petit, INRIA’s CEO, remarked: “We decided to start working on Software Heritage more than a year ago and we have now shown its feasibility. In order to scale up worldwide, the time has now come to open it up to the widest, national and international contributions.”

    Two early partners have already committed their support to Software Heritage and will help it grow. They are Microsoft, (which really needs no introduction. Ed.) and <a href="http://Two early partners have already committed their support to Software Heritage, and will help it grow: Microsoft, one of the largest software industries in the world, and DANS, an institution of the Royal Academy of the Arts and Sciences and the Netherlands Organisation for Scientific Research, dedicated to preserving and promoting sustained access to digital research data.

    Inria is now calling all stakeholders worldwide to assist the project in tasks such as, for example, helping to identify the thousands of different sites where the world’s software heritage is now spread around and contributing to the infrastructure. As regards the latter, the project’s own source code is shortly going to be released to the world and developers that share the project’s vision and want to help in this mission will be welcome.

    FSFE support

    The announcement of the project has been welcomed by the Free Software Foundation Europe (FSFE), which has released a statement of support.

    The FSFE highlights a vital reason for supporting the project, i.e. that software is prone to disappear, either because it stops being profitable, or projects get cancelled, or the code is deemed obsolete and gets erased, or is left to fade on storage that physically degrades over time.

    Originally posted on the Bristol Wireless blog.

Posts navigation