Skip Navigation Bar

U.S. Department of Health & Human Services

NLM logo

PubMed Update: MLA 2019: Transcript

My name is Marie Collins. I work at the National Library of Medicine, obviously, I'm here today with Kathi Canese. Kathi is the Program Manager for PubMed and also my supervisor.  

I am going to - hopefully in a minute or less - tell you about myself just so you will know why I am up here talking to you. I started at NLM six years ago and I spent 4 years in the MEDLARS management section of Library Operations and I've done a lot of PubMed customer service, training, and outreach. Two years ago I moved over to NCBI and now I'm on the PubMed team at NCBI. And I'm more involved now with supporting the publishers who are supplying data to PubMed, so I work on the DTD and the schema a little bit and communicating with the publishers what types of data they can supply to us, helping them get their XML in to us. I'm still doing PubMed customer service for end users and I'm supporting the people who are downloading the data from our ftp servers. So I'm sort of now really involved in the lifecycle from 'a publisher wants to supply data' to 'a consumer is ingesting that citation data.'

I want to let you know also that we are going to hold questions til the end. I have a lot of slides and a lot of material and I'm hoping to get through it with time for questions in the room today. If we don't get to questions, Kathi and I are both going to be in the booth for the rest of today and all day tomorrow barring coffee and bathroom breaks. We are both on the schedule 10-5 but we'll do our best.

We are going to record this. We are trying to record it today if the recording doesn't work out I have a date back in Bethesda to record it so that we can post it for people who didn't make it today or if you don't want to take copious notes, hopefully I'll be covering everything in the recording that I'm covering today

So, let's get started. Here's my agenda.  We are going to do the traditional PubMed update that we usually do - sort of a day in the life of PubMed.  I'm going to talk about the LinkOut consolidation; this came up during the DOCLINE session that we just finished. And then the bulk of my presentation, though, is going to be about the new PubMed that we've been building for 18 months or so, and talk about that and what's coming with PubMed Labs as it transitions to production PubMed.

OK, so if we look at PubMed today, there are 29.5 million records, there will be more than 30 million records next year. I know this because we are getting more than a million records a year. We actually are no longer OCR scanning journals and creating citation data. All of the PubMed content is supplied by publishers at this point. And over the past two years we've worked with the smaller publishers, some of the smaller societies, and helped them start creating and supplying their citation data to PubMed. So that was a learning curve for some of these journals but now we have no longer -at NLM- we no longer have to have that program for scanning and creating citation data.  We have 2.9 million unique users a day. I happened to grab these numbers in April, and even in April that has us above people who are checking for their tax return at irs.gov [light laughter].  So I just want to say, it's a lot of people. We have about 12.5 million page views a day.  And this one actually surprised me.  When we look at PubMed by device, I think we have been hovering about 38 percent for quite a while.  We are really, really, really getting a huge mobile audience. People are coming to us from their mobile devices, from their tablets, almost at about 50 percent.  So (spoiler alert) we are really focusing on the mobile presence and serving our users who are coming to us from mobile devices.

Two years ago we released our PubMed Data Management System.  As I said, our publishers are supplying the data - they are also responsible for updating and correcting the content in PubMed. When there are errors in the data most typically it's coming from the publisher because it's tricky to scrape your content from your website and submit it to PubMed without occassional errors with middle names being confused as part of the last name, and we've given the publishers the power to go in and update their citation data and this has been huge. 97% of the PubMed journals have data providers actively using PMDM. It has a very simple interface. It's just field-by-field, they can go in and edit their data.  They have made more than 85,000 manual updates or corrections in the two and a half years since we released this. Author names are still the number 1 edit.  They are also adding DOIs; they are creating links between the citation to an erratum and the original article; they are correcting affiliation data. And of course there are other programmatic ways that they can go in and update their data.  They can supply a replacement file, for example, if they need to add 300 collaborators--that's no fun to do manually.  So we do have a mechanism in place for other edits that they want to do programatically, and it's important to note that they are only updating the data that they supply to begin with.  They're not updating MeSH data, they are not fixing anything that's gone through our QA and indexing process.  The other thing is that they can request deletions but they cannot delete citations.

I want to talk for a minute about author disambiguation in PubMed because it's important to me, and I was building the slide deck, so here we are [light laughter].  This is a huge issue not just in PubMed, but in life. You will aways have one person with two different names because they changed their name, and you will always have a lot of people that share one name and authors in PubMed want to find their papers. We've done a lot of work over the past, well, six years certainly since I have been here to improve the author disambiguation in PubMed.  If you click on an author name, if you're at a citation and you know that that's your author, and you click on the author name in that citation then we will do our best to sort by that author name the publications and cluster those papers that are most likely authored by this author at the top. We've had a lot of success with this; people are clicking on these author links quite a bit.  The other thing you can do is of course, create a bibliography that you can share but this is what I am really here to plug.  Authors can register for a unique author identifier. If they supply it to the publisher, the publisher is supplying it to PubMed.  People can search on that identifier and then get all of that person's papers.  So if you want Joe Smith, you add an ORCID unique identifier to your search you will get the Joe Smith that you want to find. And the ORCID ID is part of that author name sort that I talked about in bullet number 1.  So that's why it's critical to have the ORCID ID sent through.

We currently have 770,000 records that contain ORCID author identifiers. They have several identifiers so in the database we have millions of these identifiers right now. And if you look at this chart you can see from 2017-2018 about 300,000 were supplied with the data as they came in. And if you think back to our million records a year this is about a third that are coming through with these ORCID identifiers. So please, if you know researchers, if you know anyone who is creating data or anyone who is creating data or anyone who is making articles or data available or anything like that, please, please, please, encourage them to include the ORCID ID. Publishers will send them along when they have them.  And we actually never said ORCID actually, we would take any format ID, but ORCID has turned out to be the VHS of identifiers [light laughter] we haven't seen anything else.

And if you're ever curious you can actually search "ORCID" in PubMed and you will get all the records that have these if you want to see what they look like.

OK, so new since last year, we have included a mechanism for publishers to supply formulas in the citation data in PubMed and if you're not familiar with what used to happen... well, I'll tell you everything... formulas in titles abstracts and keywords we're taking plain language summaries and they can also supply reference lists.  

So, previously, if we had an abstract in PubMed that included a formula, I don't know if you can see there, but it says formula see text. So in an abstract like this what we end up with is a very heavily redacted abstract I mean there is a lot missing there.  Now that we are allowing them to supply the MathML and put those formulas in, you can see that that enriches the citation data. And this is something, having worked with the publishers now for two years, they really are good about responding to these things.  We spam them every year and say, "Now you can supply this... " and we do have the big providers providing the content for the majority of our journals and they are able to integrate this into their systems and supply us the data whenever we change the schema.

Similarly, we are taking plain language summaries so this is an example, just to show you that if you are looking in PubMed you will find the plain language summary right below the abstract - it looks just like the abstract, it's got that structured label. And just to make it a little more readable, I grabbed the first three sentences out of that plain language summary. I'm very excited about this, because PubMed has never been designed to be primarily a consumer database.  But people who are consumers do go looking in PubMed for information - you do search PubMed and this is really fabulous and they are starting to send these data and this kind of links back to what I said about ORCID. We haven't gotten a lot of this because we just opened it up in December, but if it follows that same trend that we did with ORCID IDs I'm hoping in five years we'll just have so much more of the rich abstract data whether it is the formulas or the plain language summaries.

I want to talk about the LinkOut consolidation, this came up in the DOCLINE user's group as well. First I want to say that Kathi Canese and I will be in the booth for the rest of today and all day tomorrow, and I don't know if I see Sarah Helson in the room. Oh she's way in the back, she is the LinkOut team lead who is leading this effort.  About 5 weeks ago I joined her team as back up for some of these requests and in five weeks we've responded to 110 requests. I know I've set up 30 libraries with a new Outside Tool [audience member raises hand] Were you one of them? Thanks. Oh, Susan, of course, it was so fun.  No, we even talked in Dynamics, we were going to talk at MLA. This is great.  

So, we're working on the LinkOut consolidation but Sarah Helson, Kathi, and I will be in the booth to answer your questions if we don't get to your questions right now. In a nutshell previously we had the three flavors of LinkOut. You could supply your holdings through the Submission Utility or through LinkOut Local files and then the third option was the Outside Tool which I will talk about in just a minute. When you supply your library's holdings to us and we keep the data in the XML files on our servers, I'll just show you what that looks like.  If you have a Library LinkOut holdings file or you've supplied it to the Submission Utility, we end up with a query for all of the holdings that looks something like this. We've got an ISSN and we've got a date range and it says "For this library they have access to this range of articles." The trouble is that the program grew and we ended up with thousands of libraries with thousands of these queries that are processed every night in order for us to maintain the local holdings information.

This means every night we are indexing millions of queries in order to keep the data up to date, and it just became technically not sustainable.  We're moving to a model that supports just the Library LinkOut using Outside Tool and in this scenario, you have your library icon, as Susan can tell you, it's appearing on every citation in PubMed and when a user clicks on that icon they will get to the full text or they will get to an ILL form that is pre-filled with the citation data. Now what's nice is as, and I think that Lisa talked about this, as we make some hard choices at NLM and we think about where we are going to put our resources and where we are going to focus and how things are shifting, we are then able to move on and provide a richer service in that the Outside Tool icons will now be appearing in PubMed Labs and PubMed Labs is a truly responsive site. This means that the features that are in PubMed will be the features that are in PubMed regardless of whether you are coming from a mobile device or your desktop or your tablet. And it won't look exactly the same because sometimes you have things spelled out on your desktop that need to be made into a button on your mobile device so it won't be identical, but when you go from desktop to mobile to tablet it will be very familiar and it will be fully featured.  

So now, finally, we are at my favorite part of the presentation. So I'm not as involved in the PubMed Labs development as I'd like to be because I'm working on production PubMed. That's going to change this summer because we're sort of in the home stretch for building Labs and some of that will be building the online help and some training and doing that sort of stuff over the summer, so it's an exciting time. I'm going to start with this mission statement for new PubMed.  This was written by people in a room from across NLM from Library Operations leadership to NCBI and people from all levels, from staff on up. I love this mission statement.  I didn't come up with this, someone else did, but I love the way "connect" is in the middle of that paragraph sort of connecting the people to the world's leading sources of biomedical information. I'm not going to get into the NLM strategic plan or the NIH mission statement but I can tell you that connecting people to the world's leading sources of biomedical information is what we do. It informs everything we do. We get emails every day-- everything that is going on working in the context of NIH is about connecting people to the world's most reliable sources of biomedical information.

I'm starting with the road map because this is the question that we get. Here we are in the spring of 2019. PubMed Labs is available and it is not production PubMed. That's coming in September. In the meantime, current production PubMed is current production PubMed.  And in September when Labs becomes production PubMed, PubMed will still be archived and available. We're not going to shut one off and turn the other one on and have you lose the current production PubMed that day.  We'll have current production PubMed available for several months and if our September 2019 date slips, the January 2020 date will slip as well. They will sort of shift together.  

I have goals here, I can tell you I could do half a day on goals but I'm going to do one slide, so bear with me.  

We're moving to the cloud.  This to me is a time when we are essentially moving off the mainframe again. I think it helps to think about it in that way. We have a current environment, we have a current architecture, we have a current database and system and user interface. We're shifting everything to the cloud. It's hard to build a system with a new database, with a new UI and have it be exactly what it was that you're used to seeing in production PubMed, and it won't be identical, it can't be identical, it shouldn't be identical.  And Lisa talked about this regarding DOCLINE, that it's really a chance to say 'what do we need?' 'what can we do better?' and 'what do we need to let go of that was a hindrance?'

We always need a high-quality search.  We've moved to a SOLR backend and this gives us some out of the box features that are great. It has improved synonymy, the indexing. A lot of the features that we need are there. We've had to customize this quite a bit as well because we need our automatic term mapping and we need to turn off some of the synonymy, so it's sort of a double-edged thing where it comes with a lot of great features and we have to say "Ok, that's awesome, but our users don't always want the plural mapping to the singular or vice versa."  

It has to be fast and it has to be reliable, and everything is measurable, of course, if you try hard enough to measure it. But fast and reliable are easy to measure and so when we talk about what our goals are for PubMed we know that it has to be fast and reliable and this is one of those things where it won't become production PubMed unless it is at the same level we're at now with our current production system.  

We're looking for a more modern experience and part of that is the truly responsive website.  

And we want it to be beloved. Now, not everyone loves PubMed every day, but I think it's beloved [light laughter] I like to think. I've dedicated six years of my life to this so I hope. PubMed is beloved; people do love it. I hope some of them are in the room right now.  We're really hoping that PubMed labs will be beloved in the same way once people use it for the first time, figure out where things are and say 'Oh, ok now it's easy." The first time I heard 'beloved' I was like, "Well, that's a little hokey." I've embraced it, I really have. People love PubMed and I really hope everyone will love the new PubMed.

I'm not going to talk about agile very much. Lisa talked about agile during her talk about DOCLINE. The point that I want to make I'll talk about this - we talk about the minimum viable product which is enough features for early users so that you can go in and start playing with it and it gives us an environment to capture feedback for future development.  We're releasing new features frequently and we're paying continuous attention to technical excellence and good design.

We're releasing features more frequently as we ramped up to this date of September 2019. One thing I'll say is that if you don't see a feature in the new PubMed, don't think that it's gone forever. We have a long, long list of things that we're adding. The other thing is if you see a feature in Labs, that doesn't mean it's done. There are some features I'll point out today that are there but they are going to evolve and get better and we are enriching it all the time.

Here's our agile development model.  In an agile system you're planning, designing, building, testing, reviewing, and releasing. This is sort of the development model that has been around for ever, right? Like if anyone else worked for a government contractor in the '80s.  I don't know - raise your hand - oh, just me.  [laughter] So we have this same development model but you did it once. You wrote your requirements and you built a system and maybe you checked in once in a while but there was never a point four months in to say "Oh, what I said I wanted wasn't what I wanted. What I said I wanted wasn't usable. What I said I wanted isn't translating well in this product that you're building." And that's what agile really gives us. It gives us the chance to go in with requirements, start building and instead of doing this development model once we're doing it all the time.  We're doing it on very small pieces of code.  Well, I say small, you could build the advanced search, and you have--throughout that cycle--opportunities to get feedback and tweak it and tweak it and tweak it because you are doing your development in chunks. And then some of the tasks we have are as small as stripping a space out of a certain field. So this is scalable, this applies to a small feature and a large feature.

This next little animation I have, the little stars popping up, these are the places in this cycle where we can incorporate that user feedback. So it is very iterative, and it's very much based on user input.  It's completed in short cycles and it's increasing flexibility, improving efficiency. It reduces wasted work because you can go down a path for three days and then back out of it. You're not developing for a year and then saying "Oh no, what have I done?" like we did in the '80s.

During this process when we talked about user feeback, we've talked to researchers, professors, librarians, journalists, grant writers, biocurators, health advocates, PubMed team members, pretty much in that order, we were last.

Alright, here we go.  We have a new look. How many people here have played with PubMed Labs?  Wow, that's great.  That ends the interactive portion [laughter] I mean I wish I had postits like Lisa had, but that's it. Feel free to get water, I don't know what to tell you.  So if you look at this we have our results by year graph that people are used to. In PubMed Labs you can actually expand it so it will take over your whole screen and then you can zoom in on certain years, you can change your range. Those two little circles there you can make it change your results right there in the results by year graph.  The default sort order is best match but you can toggle to most recent. We have snippets with highlighting so when you run your search you see a little bit of the abstract there, you can kind of get the context of the result and we have filters. Filters are a feature that is in here but is not done. So don't panic if you don't see your favorite filter in the filter menu. One thing I'd like to point out is that we have this new category of "Article Attribute" which includes the associate data so you could limit your citation search to those records that will come up with associated data.   And then the little gear on the right allows you to change the format--whether you'd like to look at the summary like we're looking at here, or the abstract.

We have the cite button which PMC has had for quite a while.  And I'm very excited about this. If you click on cite the citation pops up and we have the AMA, ALA and MLA formats. We're also working on the NLM citation format, so we'll add that to that menu. You can copy it. It comes up highlighted so you just click on copy and copy it, or control-C without having to highlight with your cursor.  It's the small things, sometimes, I think. You can download it in RIS format. If you're not familiar with RIS it is very similar to the MEDLINE format that you see in PubMed except it's an industry standard so now this is something that will be really helpful for importing the citation data into your bibliographic and citation management software. We have a share button that looks like the share button on other websites. You can share to Twitter, you can share to Facebook or grab the permalink.  

There's improved navigation, so you can go backwards and forwards without going back to the results page. Personal favorite. I got ahead of myself with the animation, but let's see if I can go back. If you hover you can see what's coming next. I'm excited, we have some nice things now. We have the full text links, we have navigation on the right sidebar.  

Here are some of the features...early on with the feedback we were getting... People would come to PubMed Labs and they would use PubMed Labs and they would click on the feedback button and they would say "Well, this is great, but I just wasted four hours searching with Labs and I can't save any of my results to a file." Not like that, but kind of like that.  Now I'm at a point where I can tell people to get in there and play, and play, and play, because we do have so many of those features that make it worth your time. You can actually do something with your results.

We have the save and email and the more actions button there, which you can see is the three dots which is sometimes called a hamburger, sometimes called a snowman, I can't remember the third one, I'll think of it.I learned a lot in the past couple of weeks. If it's on its side it's a kebob. I'm just saying I might be old but I can learn new things. Save to file - it will expand right under your results and then your options will be to save all of the results on the page, all of the results from your search or just what's selected and then you have those formats to choose from.

Email is similar, because I'm logged into My NCBI account. That's my email address if you want to copy that down while you have the chance. I populated that in case you want to get back in touch with me after this. You can type in the subject and the body. The more actions menu will get you to clipboard and collections. You can add up to 500 citations to a temporary clipboard and once you've created your clipboard you'll see there is a link next to Advanced search, and it also tells you that the item is in the clipboard. From the clipboard itself while you are viewing the clipboard you can save, email, cite and share your results.

Collections, you can add up to 1,000 items to a collection in your My NCBI account. And that's pretty much the same - it's going to look the same as your collections look today. You're either creating a new collection or you are adding to an existing collection.

Ok, oh my gosh, I'm talking a lot.  I'm giving everyone a chance to take a deep breath.

Alright, so Advanced search and History. We've been showing this in the booth and think that people are having a deep sigh of relief because they hadn't seen it yet and they didn't realize that it was added to PubMed Labs. Advanced search is very similar to Advanced search now. You can search specific fields, you can show the index, you can build complex searches. This is the tool that even if the usage is not as high as some of the other features in PubMed, we understand that it's a powerful tool for advanced users. You can see here that you're able to see your field list and select a field to search in. You can view your search history, combine your searches, compare the results of various searches just like you do today. And then we've expanded it so if you go back here when you click on the little carrot, I'll show you this--I have a short video to show this--you can see your search details in the advanced search page, and that's collapsible so you don't have to look at it the whole time.

So I have a quick video because this was a lot of screen shots, so I thought we'd just show this, if it works. Oh, here we go. You can see that you can select the field that you want when you are searching. I'm keeping it here at all fields. When I type in my search terms I just add it to the search box. And then this is a split button so that you can select 'add to history' instead of 'search.' And then that will stick.  So for anyone who wants to stay in the Advanced search page and just add their items to the history without being kicked back to the results page after every search, you can stay on this page and build your search.

You can see your index list.  Just to let you know this is still there, this hasn't gone away. I'm just selecting the top option adding it to the search builder. You have all the Boolean options there. That was really fast, I don't know if you saw that, but you have all of the Boolean options.  Add it to history again, and then I think I have one more, sorry for the delay. You can tell this was sort of a seasonal search example. Anyone else from DC? In April, ya, I didn't have allergies til I moved to DC, but there they are.

And then you can use the query box also to combine your searches like you can today. So you can do that #1 AND #2 AND #3 and again add it to the history. We're building up to the best part - so like I said you can pull down that carrot and you see your search details right there, as you go up to the more complicated search it is going to give you the search details that incorporate those other three smaller searches that build up. And then I'll show you the action menu. You can add a query, delete.  Today if you say "Save to NCBI" it's not working. But in the next week or so - and it tells you, it says this feature is coming soon - but very soon, being able to save searches to My NCBI and create those alerts is pretty much next with the Outside Tool icons coming.

So we add this - three minutes gets really long when you are watching Advanced search, even if you love Advanced search more than anything. So there are your search details. But don't worry it's going to end.  You could click on results to go to the home page and see your results, there's that actions menu again, and then you can download your results like you do today and you can delete your results and be done.

PubMed Labs is a truly responsive website so mobile's going to look very familiar. You can now log into My NCBI from a mobile device, the advanced search is available in mobile, filters, timeline, selecting your sort order, those cite and share buttons, links to the full text through those publisher icons, and coming soon the links to the Outside Tool icons.

So we have advanced search on a mobile device for the first time.  Search History: putting search history on a mobile device just doesn't look great. So it's there for tablet, it's also there on your mobile phone if you're willing to flip it to a horizontal view. So if I take my little iPhone and I flip it to a horizontal view and then I scroll down I'll actually see the search history. I'm not actually sure how many people are going to use their search history on a mobile device. It seems frustrating to me but we definitely have it on a tablet and on desktop and now on mobile.

Features on the way, in this order. Outside Tool icons, saving searches and creating email alerts, that full integration with My NCBI, the ability to search PubMed from the MeSH database.  I used to call this integrated search with MeSH, but then people thought we weren't using MeSH in searches in PubMed Labs and that was just a huge misunderstanding. MeSH on the backend is integrated into Labs and all your searches are run mapping to MeSH and all of that automatic term mapping and everything is already there. What we are talking about here is going to the MeSH database, selecting your search terms and then going to PubMed to search. And the same thing with the NLM Catalog - starting your search in the NLM Catalog and then going to PubMed. And links to all the other Entrez databases, and a new PubMed API.

So I've said this already, and we have updates, whenever we have new features come out we are putting articles in the Technical Bulletin and this is actually sort of serving as a place holder for help because we are rebuilding our help system and we have a new CMS coming so that we can start building the help and enriching the help so in the meantime we are really putting some full articles in the Technical Bulletin that include screenshots and all of the directions so that people will know how to use the features and we're going to transition that into our new help environment. On the PubMed Labs home page you scroll down and it tells you what was released on what day.  

I have resources, just things that I mentioned. The slides will be available somewhere. I'm sorry, Kate Majewski, where will they be? Technical Bulletin - the slides will be available for sure, and so you'll have handy links.

So I'm so sorry I talked for so long, but we have 15 minutes for questions. Does anyone have a question? 

Question: I have a question about the Outside Tool. How does it know where to send the interlibrary loan form? Do you have to be logged in and IP authenticated?

Answer: Yes - it knows because your link resolver - we're basically taking your link, adding the PMID, and then however your link resolver is set up, so yes the person would have to be authenticated on your system whether it is the IP authentication or username and password.  Whatever your authentication is for your library. And if you have a lot of questions definitely come by the booth because Sarah Helson is in the back of the room and she's been answering questions and Kathi and I can talk to you. And even when we get we get back to NLM we can look at your account and see how you are set up now and talk more specifically about what you need.

Question: Will there be a difference in search syntax because what I saw now is that for instance phrase truncation now, when you truncate a phrase in production PubMed now it will treat it as a phrase, but that is lost in PubMed labs. Are there other differences in syntax. 

Answer: There have been so many conversations just in the past two weeks about phrase searching and truncated searching. [Kathi Canese speaking:] As Marie mentioned we are doing this in the agile way. So we're throwing features out there and we are continuing to modify them. And basically phrase searching and tagged searching will basically duplicate what's now in production PubMed. We just had a fix go in on Friday for that. There are some nice new features to the database that we are using. For instance, now in production PubMed truncation or wildcard searching truncated at 600 terms and in the new system there is no truncation so it's unlimited. So there are some enhanced features. But we're keeping most of what users can were able to do in the old system in the new system.

Follow up question: Because now if you use quotes and truncation at the same time the quotes will ignore the truncation. Will that be reinstalled now, because that was an improvement.

Answer: [Kathi Canese speaking:] It's evolving, come by the booth and show me exactly what you mean and we'll figure it out together. [Marie speaking:] So real quick, one thing that I forgot to mention is that we have this green feedback button, it's kind of little and it's there in the bottom corner. It float with you right now. Like you can't get away from it no matter what you do, it's going to stay in the bottom right corner; it's a feature, not a bug. So click on it and write to us, and actually that has an embedded URL so that goes right to our team which is Kathi, myself, and Jessica Chan, and we're responding to those questions and we're tracking the Labs questions, so be sure to click on feedback and send us your comments as well and we're responding to everything that we get.

Question: I'm going to stand up so that people can hear me. I know that you said that NLM is making some hard decisions when it comes to things and I want to know why the hard decision is making small libraries buy a link resolver if we want to keep our branding. I just don't have that kind of money, it is incredibly frustrating, I am very upset about it. I just I can't afford it, maybe my catalog is going to do it maybe not. But why is punishing small underfunded libraries the hard decision?

Answer: [Kathi Canese speaking:] We really appreciate how difficult it is when you are in a small library we understand budgets.  We were in a difficult situation where we could no longer technically sustain the holdings of 3500 libraries. It was crashing PubMed indexing far too often. And that's one of our critical mandates is to get new data in the system daily and to have it up and running for everybody everywhere. So while we completely appreciate how difficult it is for you to let your users know that you are a critical part of making their jobs easier and helpful, we could no longer do that, technically, we just could not do it. So I'd like to emphasize, and I know that this isn't true for everybody, but the bulk of the URLs that were behind your holdings information are the default publisher URLs, so that your users aren't losing access to the full text. It's IP recognition as you know, and we understand that it's a hardship for you to purchase the link resolver but even if you don't your users will still have access to those articles. [Audience member speaking:] But they don't know who bought it for them [Kathi Canese speaking:] and I appreciate, we appreciate, we all do, I don't know, I know you all are a creative bunch and there may be other ways that you can let your users know how important you are to them, it is unfortunate that we had to take away that branding from PubMed, if you can't replace it with Outside Tool, but we are hoping that we can perhaps help you or you all can get together to think about better ways to let your user community know how important you are to their daily lives. [Audience member:] Will there be grants available? [Kathi Canese:] Unfortunately, I can't address that. I don't know. I'm assuming if there is anything available NLM will let you know through the NLM Technical Bulletin. That's generally where we announce those kinds of programs.

Question: I'm thinking about what we'll need to do in terms of updating our materials and things that we share with our users. So will there be or is there already a central location for a central location for a simple rundown of what's different or will you be making tutorials we can share with them?

Answer: [Kathi:] So as Marie mentioned we are releasing features and then we are writing technical bulletins about it, and we have in our little mindspace an idea of having a technical bulletin article with: this is how you used to do it, this is how you do it now. So that is coming I don't know when it will be available because we are still adding features but I expect that before we pull the switch on the default PubMed we will have something like that for you. Anything else? 

Question: I don't have a question, I just have a comment. Now that I listened to all the ins and outs of this, I really complement this tremendous job that you've done to transfer over, so thank you for reinventing the wheel!

Answer: [Kathi:] Thank you, that's so nice. Usually we don't hear compliments so we'll take a kudo with a criticism. Although we're happy to listen to your concerns as well. So please hit the feedback button, let us know what you need, what you don't have. Thank you all for coming.

[applause]

Last Reviewed: May 14, 2019