Core Dev update time and this week we are going to change the format ever so slightly; a few community members quite rightly pointed out that the Roadmap was not being updated and not reflecting the progress being made. The Roadmap has today been updated and will now be tied to this weekly Core Dev update, the idea being to give a little more context by linking progress to Roadmap milestones. This was only hatched properly yesterday so will no doubt evolve and improve over time. Just now all links lead to TestNet 1 milestones, as this is currently the team’s main focus, however over time and as our workload becomes more evenly spread across the Roadmap this format should prove very useful for contextualising updates.
As always community feedback is welcomed.
Update from David Irvine
Now actually back at the coal face. It has taken more weeks than anticipated to get back to code. This weekend will see that happen. We have had Chris and Ben over for a couple of weeks and that was incredible, what a great pair of guys and extremely capable. I hope to see more of this for sure. We need to get the word out and the knowledge spread. It is difficult as there are so many mini tests and designs that are not clearly documented that it helps loads to let folk pick away at things and for us to show what we have done / are doing. Really helpful.
I am working on a couple of side projects that are of course SAFE related (you will get used to this, it drives the Engineers mad), I hope to have some news soon on what could be a pretty dramatic breakthrough in cryptography for us. The ability for diskless validated and secure applications is also a very interesting and potentially very simple addition. These two issues may be extremely complimentary and offer some huge improvements to networks and application provision as we know them. https://groups.google.com/forum/#!topic/maidsafe-development/Uf2rinC90K4
Niall Douglas also presented his paper to C++ Now, this is it http://arxiv.org/abs/1405.3323 and he is now also on the core dev team of Boost in addition to the admin team, a great honour, well done from us all Niall. Niall has also been formally added to the employee ownership scheme in MaidSafe, as he is a very valued key member of the core team and a really nice person, that’s important to us.
We have spoken with and agreed to contract Ciere Consulting http://ciere.com/ for some assistance in code if and when we require it, just until the community is up to speed. The team there are very much leading edge Boost developers and have a great deal of c++ expertise, particularly in domain specific languages and high performance distributed computing. We have previously worked with them as well as many others; these quick sprints are very beneficial.
A few more interviews and meetings took place, we have cancelled nearly all such events (Dublin is still on) until we have testnets up and running; the interviews are important, but launch is more important. Paige ( @ioptio ) in San Francisco is helping us out there and seems to have started with a bang, although I noticed on Vine at the bowling recently she seemed to forget her own name Thanks Paige we will do all we can to help out SF and make that a great Pod. Further than that Paige is taking up speaking slots for SAFE and spreading the word, we all owe her for this - amazing work.
We have made a start on centralising system documentation. This will evolve quickly over the next few weeks and hopefully answer all questions. We will attempt to make various levels of docs, from summary to some math proofs and test code to show results of our hypothesis where we can. Please feel free to fork and update as much as you can, this is our weakest area, apart from getting testnet1 up and running, which we can now do with the work Viv has done with the network viewer system.
Nick has had to take unexpected leave for a few days, his father is very unwell and we have told him to stay in Texas with his father as long as he needs. We wish him all the best and hope the family are coping with this situation.
Routing Test (TestNet1 - Performance)
Performance test using routing_node across the routing network has been implemented and included into the next branch. Routing Speed Test, when transmitting direct messages, across geo-scattered DigitalOcean nodes has been carried out. The comparison with TCP under the same testing environment shows the routing level speed is close to TCP and in some situations can be faster (around 40%).
Sending group messages will incur a slowdown in performance, as expected.
Maid Node NFS has been updated to wrap the Routing interface inside it (towards Client API).
Vault Manager has been extended to store pmid and anpmid keys on the network on Start new Vault request from client (Vault Manager).
We are working towards extending the network controller to allow easy network setup on multiple machines. The implementation is nearly done and will be merged to next / master after testing is complete (TestNet 1 setup tool).
We are also running simulations on routing table with close node size doubled. On positive results, we will be able to reduce group matrix complexity and network overhead in maintaining group matrix updates (TestNet 1 scalability and accuracy).
Still to do - Strike-through completed tasks in the Roadmap