Community

Magic Online Development Update:

Summary: Progress on our two long term development projects is going well and staying on schedule. The Dev, Ux,  Production and QA teams are fully staffed and have succeeded in keeping the production system healthy and growing, while keeping the long term projects on schedule. Leagues continue to be an unscheduled project as we have to keep working on infrastructure to support the large amount of user growth that is occurring in Magic Online.

New Client: The first of four design phases will complete October 9. Phase 1 will be the most lengthy phase, and will have produced a complete design for the duel scene. The deliverables for each design phase include a functional prototype, extensive wireframes and very detailed specifications. We will be bidding out the construction of the new client to external resources. Our intention is to have all four design phases complete by the middle of 2010. The community will get to see the duel scene design in some form before the end of October. We are using the User Centered Design process, so usability studies and feedback from any other use of the prototype will drive further refinements or changes to the design. But we will have achieved a complete duel scene design in a couple of weeks from now. We are currently prototyping using Silverlight, but I must stress that a final platform decision for the client has not been made. Construction time, cost, WotC’s overall digital strategy, and the availability of qualified development resources will factor in the final platform decision.

Infrastructure Update/Single Managed User Connection: This project is on schedule to complete before the end of the year. 3/4s of the dev work is complete, and half of the testing effort. We have 5 distinct application server classes currently, and 3 of 4 have been rewritten to work with a single managed connection, as well as to be maintainable and extensible. This work must be completed before the collection server can be constructed and deployed. This project will make constructing the collection server considerably easier

Collection Server: Design and prototyping are underway for the collection server. It is targeted to be in beta in Q2 and released for early Q3 2010. All out development will begin no later than January 2010. We have all the needed resources on staff to complete it on time, with some chance that work can begin before the end of the year.

Stability & Growth: The number of users and the amount of activity on the system has continued to grow steadily this year. The system is now supporting considerably more user activity than 2.5 ever did in terms of games, tournaments, drafts, trades, store deliveries, and such. A lot of work has gone into keeping the system performing well at these increased load levels while the larger infrastructure projects are being worked. We have managed to succeed in this area without slowing down the schedules for the infrastructure work. A lot of smaller scale optimizations have combined to keep the system running well under increasing load.

Leagues: Leagues are currently in the 2011 plan. If I can find a way to deliver Leagues earlier and still ensure that the system can support the anticipated growth in activity, I will do so.

The next challenge for growing system capacity after the collection server is deployed will be match/game starts. The collection server will raise the bar on how many users can enjoy being on the system significantly, but then the challenge will become validating decks for ownership and for the game format, and getting matches started for an increasing number of users. As with Escrows, the original design relies on too many database calls to resolve decklist comparisons against relatively static data (format rules and the card sets). Even the match records themselves do not need to persist in the database through their transient state, just the starts and finishes need to be in the DB. Decklist data is also stored inefficiently in the DB, with complete submissions being stored as XML rather than just deltas as sideboarding occurs during matches. This has led to too much growth in data size. Archiving has helped to date, but isn’t a long term solution.

The studio’s feature and infrastructure development capabilities continue to steadily improve as we add and ramp up new staff. Our content and game rules development is already very efficient and has delivered solid card sets and enhancements throughout 2009. The pace of development has already picked up, and will continue to do so throughout 2010.

Gordon Culp
Director – Magic: the Gathering Digital Studio
Wizards of the Coast

Hmm, thanks for the update.

I was hoping we would be further along than this, but alas...

Well-written update, Gordon.  Thanks.  Looking forward to the output of the Phase 1 UI design.  Also good to hear that user input will still be used after that point.

Thank you.

It's especially good to see discussion of specific dates, timelines, and estimated progress. 

Admittedly, some of the time estimates do not exactly make me swoon with joy.  But this post, unlike some others before it, seems to indicate that there is an overall plan, and progress against that plan is being measured.  That by itself is very welcome news.