2. Schedule

The MOLTO dependency chart only shows dependencies for WP 9 with the use cases WP 6-8 plus the dissemination WP 10. The boldfaced bits above entail that there are dependencies to the tools workpackages as well.

By the MOLTO timetable, WPs 2,4,9 (tools, ontology, req/eval) started at once. Translation tools WP3 and use case WPs 5-6 start at m7 (Varna). Patents use case WP7 has not started due to failure of partner.

By the DoW, MOLTO aims to have working prototypes on the way. So far, each partner has been providing their own demos. Progressively, there will be more need for integration, WP3 in particular will use most of the rest as components. In the best case, integration can be just plugging in APIs, with local bilateral negotiation at best between a provider and a user. But to ensure this, we must agree in time what the APIs will provide.

As suggested in the DoW text (but not spelled out in the schedule), specification/version checkpoints should be agreed more often between the tools WPs. At Varna, we get the first update of the tools and ontology workpackages. We should get together to fix times and expected contents for the remaining internal checkpoints as well. It would help to add checkpoint dates plus time dependencies to the above schedule (turn it into a Gantt chart proper --- the “Gantt chart” in the DoW is more like a PERT chart.) It also helps to be clear just what capabilities each release is planned to offer. Proposals what to insert into the project schedule are made along the way below.

Checkpoints can be constructed from the deliverables list and the milestones table.

The deliverables list implies these checkpoints with implications to the evaluation timetable:

• M03 Molto web service, first version – we got the phrasebook running from MOLTO website.
• M12 GF grammar compiler API – what does this add?,
• M18 Grammar IDE, Ontology-GF interface, Translation tools API – a web translation platform that allows on the fly extension of lexicon and grammars with ontology tools. Evaluation of lexicon and grammar extension can start.
• M24 Translation tool prototype running – means translator tool evaluation can start with test users (on the museum and if relevant the math case).
• M30 GF tools integrated with SMT tools – evaluating the combined system with the patent use case can start. Manuals done: testing with new users possible

Milestone MS3 may need updating relative to the deliverables list. No important deliverables are scheduled between M6 and M12 that would motivate a demonstrator there. A more appropriate place for the next version of translation tool (after the Phrasebook) is after M18 . M18 should make available ontology interoperability, and along with that, new lexical tools.

Having fixed the schedule some, we go through the WP 9.1 tasks boldfaced from the WP9 statement of purpose.

[From DoW] The work will start with collecting user requirements for the grammar development IDE (WP2), translation tools (WP3), and the use cases (WP6-8). We will define the evaluation criteria and schedule in synchrony with the WP plans (D9.1). We will define and collect corpora including diagnostic and evaluation sets, the former, to improve translation quality on the way, and the latter to evaluate final results.