Table of Contents

Date

27 Sep 2016

Attendees

Actions items from previous meeting

• Peter Monks - schedule Lawrence Miller (Deactivated) to present on OSS plan in next week's ESCo meeting
• Michael Harmon, john.stecher@barclays.com, Lawrence Miller (Deactivated), Former user (Deleted): review Symphony Foundation Contribution Lifecycle Proposal and prepare comments / feedback for discussion in a future ESCo meeting
  • Peter Monks: schedule that discussion into future ESCo meeting agenda
• Peter Monks: schedule "WG ontology" discussion into future ESCo meeting agenda

Agenda

Meeting notes

• Roll call:
  
  • We have quorum
  • John indicated he'd be 15 minutes late
• Financial Objects Standardisation WG update (Bruce):
  • The WG kicked off in March and has been meeting regularly (bi-weekly).  Quite a few members, but attendance is a smaller group.  We get more participation from the partners than the customers.  We get quite regular attendance from the Cusip and FactSet folks, on and off from banks & customer-side organisations.  Focus has been to a) collaborate / discuss / share the implementation & infrastructure to support financial objects as part of the Symphony platform and b) partly to agree on specific formats for specific financial objects.  It's been a little bit of a struggle to get engagement form the other participants.  Meetings generally consist of me talking and others listening.
  • We have a wiki page for MessageML and EntityML XML.  We talk about a number of use cases, but:
    • Bank side were interested about some of these use cases, but we didn't get very much fleshed out
    • We identified a set of financial objects, but we haven't done very much in a tangible form with them.
  • I think there's a bit of a conceptual gap.  I'm hoping that when Symphony is able to deliver a working implementation in production (we have them for demo and dev purposes - 1 or 2 partners have shown interest), people will start playing with it and get interested in doing things with it.  But in order to drive through specific object definitions the question I've been asking for several months is "who is actually interested in sending actual objects to another organisation and transact another business transaction".  The answer has been deafening silence.
  • We've talked about a demonstration with 2 partners to demonstrate the cross-enterprise capabilities.  FactSet agreed to write a bot that would listen to conversation in chatrooms, listening for cashtags and then use the FactSet back-end in order to look up the string the user has typed into the cashtag and inject into the room a message containing a structured complex object with ISIN, CUSIP and whatever other useful information they have, then a second participant would provide a renderer to render that complex object in a rich way.  We have a not-very-pretty, but functional renderer developed by LLC, but we haven't gotten to the point of having another participant develop a renderer - we haven't had engagement form 2 parties to make this a reality.  Everyone seems content with the format and structure we're providing.
  • We're in a bit of a hiatus at the moment.  My hope is when we have the capability to send & display complex objects we'll start to get a bit more positive engagement.
  • And that's where we are.
  • <Q&A commenced>
  • Lawrence: what's been the best, and what are the biggest challenges?
    • Bruce: most useful has been the ability to share with potential users the technical details of what w'ere going to provide and to get direct feedback from people who are implementing that capability.
    • Biggest struggle has been getting engagement from people.  I anticipated that people would be biting my arm off "when can we have this - I want to indications of interest with my colleague Joe in that company over there tomorrow so please deliver this stuff".  There hasn't been any of that, and in fact in order to get the implementation done it's a prioritisation decision vs other customers features and  a couple of times I've said "if actual customers are actually asking for this I can make it happen" and even with that I haven't had more participation from members.  There doesn't seem to be urgency.
  • Lawrence: do people have the technical ability to move ahead in implementing this?  Is the barrier desire or technology or prioritisation?
    • Bruce: the platform team has a public environment called "Nexus". For several months now we've been able to provide an environment where pople can code against the API and send messages containing complex objects.  We've had a demo renderer which Glenn wrote, which produces a not-very-pretty but functional rendering with a graph that grows over time and a ticking price (using Yahoo public data).  We've certainly got the technical things that you need in order to be able to try this out and the FactSet team were able to build something against that and we've seen some evidence of that.  It's certainly the case that the tools on the Symphony side are available.
  • Lawrence: do we have the ability for people to build parsers yet?
  • Bruce: a parser for what?
  • Lawrence: for a user to enter these objects.
  • Bruce: in sprint 42 we hope to be able to ingest complex objects via API and render them on the web UI only.  Content export and mobile UI will not have rendering support, but that would allow you to be able to send messages and see them in the Web UI.  The ability to compose them on the keyboard is not part of the sprint 42 deliverable.
  • Lawrence: it seems to me that in viewing this as a bot → user interaction that we're enabling chickens and expecting to see eggs.
  • Bruce: I keep phrasing this as a prioritisation thing and so far the product management process favours things customers are asking for, so other features have been prioritised.  My expectation is that when we get to the point that messages can be injected using the agent API and they're rendered in realtime - when people see those things they'll say "how can I send one of those?".
  • James: I think that's right.  On my side we have a team who are intellectually interested in this.  But we're still rolling out Symphony and right now we don't have demand from our business people.  But on the flip side, the more shiny stuff we can show the business folks the better it is for Symphony.  "Priming the pump" to enable those conversations.
  • Bruce: I've discussed this with Emmet and similar people at 2 other banks who were enthusiastic about it.  I think "intellectually interested" sums it up.
  • James: the flip side of this - this is more of a product management thing - I'm getting a lot of feedback that "this is just another message system".
  • Lawrence: agree.  How do we get that feedback represented to the product management and prioritisation process?
  • James: it's a vision thing - someone needs to have vision.  Asking for customer feedback is just horses - need to wow them a bit.
  • Lawrence: I'd hope the WG would articulate that.
  • James: the WG people probably aren't the business people. 
  • Bruce: fair comment, but it's also the case that when I've said "is anyone interested in using this", the answer is silence.  The people on the call aren't gung-ho about this, they're just not that enthused.  I think this is a vision thing - it's the most important thing Symphony can do (in my opinion), and it's something that no other chat system can do.  I have argued as forcefully as I can that this should be a priority, but I'm consistently confronted with the argument that this isn't what the customers are asking for.
  • James: yeah I can completely see that.  This is possibly a non-Foundation question that we should talk about: How do we get the vision people in the organisations together to talk about using Symphony?  So we can influence where it ought to be, rather than the minutiae of "do we want a red one or a blue one?".
  • Lawrence: I think this is both.  I'm 100% in the vision camp you're making.  But I also think it undercuts the point of articulating the strategic vision when it's provable that people in the Foundation WG aren't interested in implementing this.  This undercuts the argument that we should be building more of this.  "Undermines" is a better word.
  • Bruce: I'm not sure I'd agree with that conclusion.  I think when people see something more like a trading system than a chat system from the 1990s, they'll start to think about what else they can do.
  • James: let me tell you how this has gone in my organisation. 2 years ago it was: "yay we've signed up for Symphony" and we had all these new ideas and sent out list after list to Symphony.  4 product people later we're sick of sending those lists.  We've also been cut to the bone, so we're not going to get internal buy in or resources to work on this.  A re-injection of excitement would be a good thing.
  • Lawrence: question is: how do we do that?  It's not like "if we build it they will come".  The WG may have a lack of interest, even an artificial lack of interest.
  • Bruce: to be fair I don't think the prioritisation decisions have been made based on this lack of interest.  I think it's just that we're paying more attention to the customers who want a faster horse or a blue one.
  • Lawrence: I think we all agree, but the question is how do we move ahead?
  • Bruce: I think that something the Foundation could do perhaps to greater effect than attempting to manage a code base that's not yet open source, would be to present a coherent voice of what Symphony should be, and what would excite people generally.  Maybe the Foundation is a forum through which we orchestrate a coherent assertion that something visionary is something we should be looking for.
  • Lawrence: how can the WG push for that?
  • James: we can go nuclear: either you all put in requests to your PM or alternatively we close down the WG as being pointless.
  • Lawrence: do you think that'd work?  
  • Bruce: it'd free up a lot of time every 2 weeks on everyone's calendars!  
  • Lawrence: that sounds like a "it wouldn't result in work" approach.  Would the upcoming hackathons be safe places where we could drive for creation of proofs of concept and trigger people to bump this up in priority more explicitly?
  • Bruce: we could certainly try.  I also didn't mention that Symphony is working on a JIRA integration (being announced at Innovate) where JIRA events will be triggered in Symphony.  It's using the entity markup that was invented to support financial objects.  This will showcase the kinds of things that the platform will be capable of.  But I think using the hackathon is a great idea.
  • Frank: inline representation of content is absolutely crucial but what we're going to present at Innovate from the perspective of a trade break being injected into a message, with the transaction coming back out of Symphony to our back end system.  This is a similar use case to what we're talking about here (transactions injected into conversation, with context).  I'm all for that, and I could demo that at the dev conference if you'd like?  Happy to show where we need to be going from a financial marketplace implementation.  This WG has tried to move in that direction, but the tools aren't there yet to support what Bruce has in mind.  This also isn't new - we sent some info to Symphony early on on inline content, and it was ignored.  Some of these ideas need to come back and Symphony needs to take them seriously.
  • Lawrence: are these missing APIs or lack of integration features?
  • Frank: both.  Rich content injection via APIs, adapters / features in Symphony to get content inline from conversations are both crucial.  To be clear, what Bruce is presenting in the POC is exactly or very much what we need to see supported, along with the APIs that let us use those features.
  • Bruce: <talked through the following screenshots>
    • <mockups of the objects and flow>
      • 
      • 
      • 
    • <screenshots of POC code>
      
      • 
    • Bruce: It's developed a little bit since then.  The object can be opened and closed, like a Youtube video.  In fact we plan on replacing the current hand-crafted inline content renderers for hashtags, cashtags, videos etc. to use this framework.
  • Frank: that's a great segue to the MTM implementation that we've implemented
    • MTM is Market Trade Manager - handles push trade derivative clearing.  This system provides the ability to rectify a dispute.
      • 
    • Popup in Symphony UI showing the dispute:
      • 
    • In this example the trade dates are bad (different).
      • 
    • The Markit system is picking up these actions from the conversation.
    • I'd want to see the table above in a better format - include some information from the derivatives space.
    • Lawrence: and highlight the differences?
    • Frank: exactly.  This is a limitation of the current API we have - we can't create rich content in tables with the MessageML that's there.  The markup it provides is very limited.
    • Lawrence: this is pure messageML, not an iframe?
    • Frank: correct.  We want to inject more lively content inline, to capture the interaction between the multi-party conversation (bot, 2 counter-parties).
    • Lawrence: I see the points you're making about limitations.  I'd like to follow up with you offline, and with Michael Harmon.
    • Frank: we'll be demoing this at Innovate.
    • Lawrence: it's a great example of where we need to head.
    • James: what would be useful would be to draw out these limitations around basic grid rendering.  This is only a small improvement over structured text in Bloomberg - it's not the "step up" we'd hope for.
    • Frank: Hopefully this helps supports what Bruce has been explaining.
    • Lawrence: super helpful.
  • Peter: Nathan Bricklin (Wells Fargo's SSF board member) has expressed interest in forming a WG to discuss messaging strategy rather than technical details - some of the earlier conversation seems like it might be aligned with their request
    • John: Same topic here within GS: "when do you use what?" is a constant question.  What interactions occur inside vs outside? etc.
    • Lawrence: the LLC needs to be involved in these conversations for them to be meaningful. I'm sure those conversations are already happening internally.
    • John: that's great but this is more broadly in terms of us wanting to communicate with our customers and counter parties
    • Lawrence: the LLC board has been active in this realm.  The question is how can we fit these conversations together?  Let's start with the connections, then see how an ongoing WG-type of engagement would work.  WG's need to be self-sustaining.
    • John: agreed - the WG has to continue to provide value.  The compliance / regulatory front is a continual battle that I think should be part of this.
    • Lawrence: I think there are conversations going on and we just need to align them.
• <out of time, meeting adjourned>
  • Reminder: no ESCo meeting next week (Oct 4th)!  We will resume the regular schedule on Oct 11th.

Action items

• Lawrence Miller (Deactivated): follow up offline with Former user (Deleted) and Michael Harmon regarding limitations around rich content
• Lawrence Miller (Deactivated): locate internal LLC conversations (if any) around "messaging strategy and vision", and then work with john.stecher@barclays.com and Nathan Bricklin on potential new WG on this topic
• Peter Monks: push agenda items we didn't get to to subsequent meeting(s)