MT-Archive | News Archive | MTNI

NEWS ARCHIVE

Public Tender Invitation

October 31, 2017

The European Association for Machine Translation (hereafter EAMT), on behalf of the International Association for Machine Translation (hereafter IAMT), invites anyone in the public to submit a tender in relation to the provision of the following services for the IAMT.

Description of services: Vendors are required to provide the following services:

LOT 1: Conversion of the information currently available in www.mt-archive.info into a database format allowing easy maintenance of the information, and setting up of a new server offering at least the same functionality as the current www.mt-archive.info to the public through that database. Please refer to Appendix 1 for detailed specifications.

LOT 2: Annual maintenance of the new MT Archive server, including hosting and name services. Please refer to Appendix 2 for detailed specifications.

Tenders for LOT2 are subject to submission of a tender for LOT1.

Validity: The tender shall be valid for at least 6 months from the date of the tender.

Eligibility Requirements: Any individual or company capable to provide the above services is eligible for consideration of their tender. For the EAMT to assess this capability, the tenderer should attach a declaration providing any information they deem appropriate about their qualifications and previous experience in a brief 1-page document attached to the letter. Any decision taken by the EAMT about the eligibility of a tenderer will be final.

Submission of Tender: Tenderers have to prepare a letter including the following information:

  • The name and full contact address of the person responsible for the tender.
  • The service lot(s) for which tenders are offered.
  • The date in which the tenderer will be able to start offering the service.
  • A technical description (complete but no longer than 2 pages) of how each service lot tendered for will be implemented.
  • The total price in US$ for each lot.
  • A 1-page declaration of capability as described above.
These documents should be submitted as scanned, signed, PDF documents in an e-mail message to EAMT Secretary Lucia Specia l...@dcs.shef.ac.uk

with subject line "Confidential – Tender Document for Provision of MT Archive Services for the IAMT" on or before December 15, 2017. LATE SUBMISSIONS WILL NOT BE CONSIDERED. Please note that it is the responsibility of the tenderer to study the specification and requirements as given below before submitting the tender.

Additional information: Should you require more information, or for any other queries, please contact EAMT President Mikel L. Forcada at e-mail address m...@dlsi.ua.es.

APPENDIX 1

Description of Lot 1: Conversion of the information currently available in www.mt-archive.info

Server www.mt-archive.info was created by John Hutchins in 2004 and currently contains about 12,000 entries. All of the archive (various portals and indexes) is hand-crafted (HTML pages are written using Microsoft Word), and all of the papers are stored as PDF files. It is the single most important source of papers about machine translation, with emphasis on historical MT papers.

The tenderer is requested to offer the following services:

  • Setting up an alternative server (including hosting and name services for 1 year).
  • Mirroring, parsing and ingesting the largest possible fraction of the information provided in the HTML files of www.mt-archive.info into a database. About 95% of the entries are expected to be parsed and ingested.
  • Providing a way to reproduce, from the database information, a copy of www.mt-archive.info, using literal HTML for those portions that could not be ingested into the database if necessary.
  • Providing a new, intuitive interface allowing users to search for any paper currently present in www.mt-archive.info through any of the criteria currently available in www.mt-archive.info, and with a functionality similar to DBLP.
  • Creating an interface to easily add new entries to the database for new content created after the migration; alternatively, make available a simple API to do this remotely.
  • Papers should have the current short URLs they have in www.mt-archive.info so that they can be linked. Their short names are currently unique and could be used as keys.
  • The system should be capable of generating complete, accurate BibTeX records for papers (note that automatic BibTeX entries generated by Google are usually rather incomplete).
  • Managing the migration so that the new name of the server is www.mt-archive.info.

Options that will be considered:

  • Option 1: It should be possible to search papers by contents where the text is easily extracted from the PDF files. As many historical papers are not only scanned but OCR'd and corrected, their contents could also be indexed and searchable.

Conditions:

All of the software used has to be free/open-source software, and the server must run a common distribution of GNU/Linux. A LAMP configuration (Linux, Apache, MySQL, PHP) or a functionally equivalent configuration is expected.

All of the software written has to be provided, with source code documented with sufficient detail for providers of Lot 2 to operate and debug the server, and under the GNU GPL v3 license, to the IAMT. Ownership of the software rests with the IAMT. The code has to be delivered and kept in a revision control system (Subversion or GIT).

The maximum price for this lot is US$2,000.00.

APPENDIX 2

Description of Lot 2: Annual maintenance of the new MT Archive server

The tenderer is requested to offer the following services:

  • Setting up an alternative server (including hosting and name services) for 1 year after the completion of the work in Lot 1.
  • Backing up the contents of the server.
  • Ensuring 99.5% availability of the service, as described in Lot 1, with downtimes never longer than 48 h.
  • Debugging the code run by the server so that the continuity of the service, its quality, and the user experience, are consistent; in particular:
    • The search interface for users.
    • The database maintenance interface
    • The database maintenance API
Keeping the server up to date as regards the papers published in all conferences dealing with machine translation listed in http://www.mt-archive.info/srch/conferences-1.htm.

Conditions:

  • The maximum price will be US$1000 per year.
  • Updates will be pushed or committed to the revision control system established in Lot 1.
  • Ownership will rest with the IAMT.