 Brought to you by MarkLogic Now here are your hosts Jeff Frick and Jeff Kelly Welcome back everybody. We're live here in the Silicon Valley area at MarkLogic 2015 our logic world I'm Jeff Kelly with Wikibon. I'm here with my co-host for the day Jeff Frick And we're joined with by our next guest. Her name is Beverly Jamison. She is the senior director IT architecture and publishing solutions at the American Psychological Association. That's a mouthful. We're happy to have you here Thank you. Pleased to be here. So Jeff and I were looking into your background a little bit before the interview And we see that you your background is in graph theory and classic Combinatorics. Now Jeff and I looked at each other. We said first question got to understand what is combinatorics? Graph theory is actually a portion of the field of combinatorics. It's the larger field of combinations, permutations and a lot of discrete math on which some of computer science is based Well, so let's talk a little bit about your organization and what your role there is tell us a little bit about The APA kind of its mission its core charter and kind of what you do there American Psychological Association is the trade association for psychologists So they perform the typical functions of a membership organization with roughly 140,000 Psychologists in it in addition to that APA is an academic publisher That's actually probably its major business unit. We produce the psych info database It's like Medline that covers the field of psychology and we publish books and journals psychological tests and measures videos and the like Psych info database for example has about four million records in it So we we deal with large quantities of data Supporting that the technology that provides all that content to your members Everything from the production of it to business to business interchanges to the delivery to our end users So tell us a little bit about your relationship with Mark logic You know you're here obviously at Mark logic world I think you were just giving a talk tell us a little bit about some of the things you're doing with Mark logic and how you're using No sequel technology to support some of those things. You just talked about sure we We started out as kind of a classic case for the use of Mark logic back when it was referred to as an XML database kind of before no sequel Movement of today got started. We were a media company. We had large quantities of XML We had just gone through the big transform from SGML to XML But we had different pipelines all over the place coming through so Mark logic was a great opportunity To use a native XML database the relational databases were putting XML in but it was kind of early days It wasn't scaling whereas Mark logic was built for it So it was a very natural fit and we also kind of knew what to do with it and get all our production to terminate there And we grew from that to using it for business to business exchange and then finally Moving our search into Mark logic So you've seen the kind of the evolution of Mark logic over a number of years How would you characterize kind of the direction they're moving is it more to a full-featured? Kind of approach database platform versus kind of a more specific niche tool that it was maybe a few years back That was kind of exactly our experience and it coincided well with our our needs Once when we moved our search into Mark logic the biggest gain that we saw actually was in our data flow We cut 40% off our time to market because we produced the data in the same place that we delivered it And it was a fast fast move from there, but it also meant that kind of all the angles of that publishing business were flowing through Mark logic and Then the then our public website decided they wanted all those search features too So now we had multiple websites using Mark logic and all of a sudden it was in a different place in our infrastructure instead of being There for a very specific function It was now the core place for data services to happen and it meant we needed to think about it differently And fortunately for us Mark logic was also thinking about it differently And how was that value delivered to to the users of the data? What were you seeing in terms of their behavior and their feedback and what they were doing that they couldn't do before they could do better? Because the first thing people noticed just was the data of getting to them more quickly But also live alerts which had been a little bit of a specialty feature before now Anyone whether they were students or academic users or members in the member services area For example could sign up for alerts either ones We had prepackaged or anything that they wanted to create so people noticed that and then In some ways some of the features were things that People didn't notice then as they they wouldn't notice that you could take all kinds of Related data act as if it were the same and your type ahead still worked in your highlighting still worked So some of the changes just everything worked the way that you did the increased speed and access drive an increase in consumption Yes, it did when we first launched that search The feedback all around was it was zippy That helps and then on our public website it also allowed us to be more of a communication channel Okay, we were we were able to add features like What we call best bets or APA recommends so someone could throw a term at us. We could show them Things that we thought they might want to see Documents that we knew were of some importance in the field. We could at the say we could show that on the same page We could adjust our relevance algorithms to to tune for them So a lot of those features I think just it helped people find things, but it also helped us communicate things to people any big surprises any positive unintended consequences of suddenly this improved performance Well, of course one of the Things you get with improved performances. They always say in technology the only thing that grows, you know faster than More storage more speed is the demand for more Nothing is one more right? We learned how to do clustering and scaling really quickly because more applications went on to to mark logic and Usage went up. Yeah, we'll talk about you We hear it hit the conference here in mark logic's marketing around there the only enterprise no single database So, you know when you talk about the term enterprise, you know, what does that mean? What does enterprise grade really mean to a practitioner from your perspective? You know, certainly as you start to use mark logic for a number of use cases Kind of scaling that out clearly it became critical part of your business And I'm guessing whatever enterprise grade means to you was important to you that it What are some of those things that our logic provides from an enterprise grade capability That allows you to have confidence in it to support so many wide-ranging use cases The clustering and the configuration management came a couple versions ago and they were greatly Desired because all of a sudden many different applications needed to interact with mark logic And if we made a change we just we needed to test it You know much more majorly way and we needed to be able to deploy potentially to multiple environments. We needed to scale up the DR capabilities and We were also particularly grateful for the extremely smooth DR swap over Part of that's the way the clustering moved also part of it is just the The code is in what they call their modules database. So deploying your code is transactional just as Deploying data is so you did not have to worry about what's going to happen to your files while you're deploying code Old stuff's there this stuff's there, you know end of story and That's good for deployments, but it is also really good for for DR So another you know important topic here, of course semantics Talk a little bit about that and what how that's translated at your organization to actually provide real value to your two members Semantics in part allow us to diversify how we look at the data. We started out pure x-query and Of course in the no-sequel world. We headed to to Jason and With the RDF triples It means that we can look at the data as relationships as you kind of expected the American Psychological Association Relationships are important things or you can think of relationships between people entities But also between a condition and its treatment. I mean just a lot of things where we want to capture that information So in that sense it allows us to model what's going on in a more direct way But in addition to that semantics brings types of computation it's a type of data representation, but then sparkle is also a way to reason on it and That allowed us to do Some things that made our business pretty happy pretty quick a simple one APA produces guidelines in psychology and at one point they they had one on diversity I Think minorities and the like and they wanted to know well did this have much impact? We could do the traditional thing we could get a citation count or maybe a download count from the website But with semantics a push of the button and I could take that guideline and say not only who cited it But who cited them and who cited them and you could just watch how it permeated the field You can see some papers will kind of go a certain distance and then they've hit their max others You can just see them We could show Our CEO and his board really quickly what happened That's great, and then you just came out of a session before we got you on So what was your session on for the people that weren't able to attend the session was on the dramatic and glamorous topic of continuous integration Which is a means of making sure when you change one thing that you didn't break something else Again, it's a reflection of the extent to which mark logic has just permeated our infrastructure that we need to be able to Build a robust application We also need to be able to have mark logic as part of a multi-tiered application Taking advantage of what pretty much everybody's doing these days with html5 stack We want to have angular sitting there on the user's browser or you know on their laptop or their phone We use node in the middle use mark logic on the back end We want a very smooth flow, and we also want things modular so that we can plug one component Into another and continuous integration Provides the discipline to make sure that each piece is tested when a change is made Not you know not the least of which is of course you can set this up so that if anybody broke Anything and everybody gets an email right away Okay, well we've got just a couple minutes left So I want to give you opportunity to talk a little bit about what's on your road map I mean when you look at what you know mark logic is doing in terms of developing the database and new features I'm sure that's got a lot of ideas going on in your head about what we can do next What's kind of to the extent that you can share what's kind of top-of-line for you And one of the things you're gonna be working on the next six to all month One of the things is greater Visual presentations of our data Nice as it is to have search and facets Notice right now for example You can search on an author and get a nice list of their publications But of course in today's world what people would love to see is Picture of the author yeah exactly so we we have a lot of those coming You know down the pike whether it's people or topics or organizations Another one is the ability to build much more complex tools We've we've done a little bit in the area of providing tools for practitioners or educators or students But we're we're trying to do a lot more than content is helpful, but It's a more active world and then what about mobile are you seeing a big impact on mobile in your world or is that? Not necessarily hit this this group that hard yet. It has I mean that's the reason for wanting the HTML stack And it's the reason that much as I love semantics the thing that I was the most excited about in Mark logic 8 was angular node json being native to it because yes you You will be blasted out of the water if you say you're going to put something out there that it doesn't work well on a Yeah, right Well some great insights Beverly Jamison from the American Psychological Association. Thanks for joining us on the queue Appreciate it. Hopefully we'll see you back here next year at Mark logic world Thanks for watching stick around. We'll be right back live here at Mark logic world 2015 Right after this quick break