Herding Tigers – How Salesforce Let Go of Inline Links

[00:00:01.030] - Mysti Berry

I am Mysti Berry, principal technical writer at Salesforce.com. And I've been a technical writer for a couple of decades. I was a content strategist for two years and then I moved back to the Internal Doc team. And please don't hold Salesforce responsible for anything I say.

[00:00:20.450] - Mysti Berry

So Put a Link On It. That's what I did for two decades, if you're a little bit worried that a customer might not be able to navigate to some content, you just put a link and some other content and everything's great, right? No problem. Well, it turns out there are a few problems with that approach, and I'll talk about what we did at Salesforce to remove a whole bunch of unnecessary Inline Links, including how we decided they were unnecessary, how the team responded to that large project, and then some best practices if you're thinking about purging links.

[00:00:55.690] - Mysti Berry

But first, of course, I have to convince you that there's something wrong with Inline Links, and when I say Inline Links, I mean links that the writer puts in by hand in the middle of other prose content, not See Also Links or things that are generated by your publishing system.

[00:01:14.020] - Mysti Berry

So it turns out links actually do slow down every reader just a little bit and they slightly decrease the comprehension as measured by recall just a tiny little bit. And if that was the only thing that was wrong, we'd probably just live with that. But then we discovered through ad hoc testing (it isn't complete yet), but our ad hoc testing, as well as ad hoc testing at VMware, found that the links are clicked between never and eight percent of the time.

[00:01:42.470] - Mysti Berry

So we put a little bit of effort into putting in these resources and they're almost never used. We discovered, even more shockingly, that when a topic is nothing but links and the links are clicked only twenty to thirty-five percent of the time, the rest of the time the customer just goes back and searches and finds something else. And more research is needed here. But we were shocked at how low these numbers were. The resource manager, the thing that helps you build a link if you're working in XML, that can be very painful in some tools and we found that our writers were taking up to a couple of minutes to add a link.

[00:02:19.980] - Mysti Berry

And then when you have to go back and reorganize your content, fixing those links can be a real nightmare. I know some CCMS programs handle that for you or if you're in FrameMaker it's a little bit easier, but you do have to do this manual work whenever you do a big restructuring of your content. And if you're working, if you're trying to move existing content to mobile, for example, now where far fewer of your topics maybe be used, it can be so difficult that your writers are actually saying, oh, just let us write from scratch.

[00:02:53.400] - Mysti Berry

This is so bad we don't want to deal with it. And also, if we put effort into manual Inline Links and nothing else, we're not looking for better solutions for our customers. And we know from the preliminary numbers that our customers are just not using the links we're creating, even though we're spending time and money to build them and maintain them. So at our shop, which is XML Open Toolkit DITA, we found that our doc builds were actually slower because our writers have to get a clean build before they check-in.

[00:03:29.930] - Mysti Berry

And we found that those doc builds were slower in part because of Inline Links, and that's because topic A, calls topic B, that calls topic Z. And so instead of just building one book to test, you're building five or ten. I found when I did my initial analysis that there were a shocking number of circular references, probably because content changed over time and they didn't start out as circular, but they ended up as circular after just a year or two, and that topics were sometimes very disjointed because the writer was...

[00:04:04.880] - Mysti Berry

The writer knew the content at the other end of the link, but you couldn't tell from reading it. And many writers were just not doing refactoring that was needed because what seemed like a four-hour task was turning into a three-day job to get it all fixed. And this is a content that I wrote, I'll take full responsibility, I was a huge Links advocate in my early days at Salesforce, but when you sort of look at this, you can see the content at the bottom.

[00:04:34.150] - Mysti Berry

It's really hard to read because of all those links. And many of them are useless, like every link that goes to account partner, I think is actually linked to this very topic, which is crazy. And if you look at the Field Properties column, it's very unlikely that someone really needs to click through. At the moment, they're examining this topic to find out about field properties. They probably don't need to do that and almost no one will ever need to look at Boolean.

[00:05:05.380] - Mysti Berry

We pretty much all know what Boolean is. The other problem is the center of information for all these links grows very faint and people give up. And my example here is the top part of the screen is a topic about enabling a feature called Email-to-Case. Doesn't matter what it is, but you'll see it. Step three is the link oh, you have to go to this topic to figure out how to turn on the feature itself. Well, OK, that's not so bad.

[00:05:33.520] - Mysti Berry

But then when you get to the topic, you do a few things like, oh, I got to go to this third topic to do some configuring, and that third topic is just a reference topic. And now you're lost. You can't remember what you're supposed to be doing. It's two steps back and you're completely taken out of the metaphor of completing a task and you're lost in this reference content. And we know from coupons and things that, that this was true for some of our customers.

[00:05:58.780] - Mysti Berry

But also we weren't doing this round robin and this was done by the original writer because some of this content got used in many places. And so it just seemed natural to link, instead of copy stuff in. But if we don't rely on links, then we can take a step back and say what is the customer's path through the product and therefore through the content? And we can write much simpler, slightly longer in a single topic, but much simpler, easier to follow instructions.

[00:06:29.580] - Mysti Berry

And I'll talk more about how to do this later. So what did we actually do? We had an opportunity, a now or never opportunity to attack our whole health system at once with a bunch of our writers over 4000 topics and we needed to carve it. We had one giant 4200-page health PDF and of course also topics in a portal. But we were getting teased pretty hard about that giant PDF's. All right, we're going to break it up into these 13 logical bundles.

[00:07:02.490] - Mysti Berry

And we couldn't do that because the links would break and there were a million of them. And there certainly there were certainly technical workarounds. But at the time we said a link to another we called it Bundle just to try to avoid thinking in terms of books. But a link that far outside the use case is probably not going to be helpful. And it's better to get rid of them. And you'll notice in the results here, we got rid of only half the links, but that was forty-three hundred Inline Links or a slightly over more than one from a help topic, which is kind of crazy that we had so many.

[00:07:40.990] - Mysti Berry

Seems like it should be easy, but it really wasn't. We had hundreds of DITA Maps, thousands of topics and completely ad hoc, no pattern to it, no way to run a script to look for thousands of these manual Inline Links. So we got a spreadsheet. We got all our writers, our writers, whose efforts are great. They work cooperatively really well. And so we carved up, sets of topics and just went through and flagged the unnecessary, the Inline Links. And the project was so big that we actually hit the user limits on Google Docs.

[00:08:14.670] - Mysti Berry

There was so many of us in their editing docs that Google crashed on us, which I thought was kind of funny. But as we looked at each link, we had to decide whether we could just delete the link because it just wasn't necessary or if we had to do something. And I'll explain what that something is later. So there was that initial sort and that's how then we found the ones that were just obviously not necessary and could easily delete them.

[00:08:41.280] - Mysti Berry

We did that. Then we had this whole issue of necessary XREFs or Inline Links. And we decided it was necessary if in the writer's judgment and it was always in the writers judgment that readers would suffer loss of corruption of data if they didn't follow the link. And this was around importing data, and some other things like that. Or if the reader was obviously we just simply couldn't complete a task without the link and for example, parent topic with links to child topics with lots of steps.

[00:09:18.930] - Mysti Berry

And if the link met either of these two criteria, we still would just leave it in if it didn't cross the bundle boundary, mostly because we needed a way to scope the project, we knew we couldn't fix every link across the whole dataset, across a whole set of topics. And I would always ask the writer if this XREF is necessary, why isn't the content it points to also necessary? And that's when we start to have a conversation about structure and architecture and what we were really trying to do for our readers.

[00:09:52.880] - Mysti Berry

So for those things that we felt we had to address and couldn't just get rid of the link. The number one solution was just get them back into the.. Get the customer back into the UI so they're on a page and then click the help for this link or they've done a search and they want to do a task. You get them to the next place in the UI where they need to go. And then send them to the UI, then they'll either do their work or if they don't understand that UI page, you know, that Web page that they're trying to do their work on, they'll click another link.

[00:10:25.000] - Mysti Berry

So instead of navigating customers from content to content to content, here's some content. Now get back to work. Oh, you're still confused. Get back to work. Here's some more information now. You can get back to work. And that solved, I don't know, maybe 30 percent of those necessarily questions.  if you're pointing to another topic just for a small piece of information. Just go ahead and Conref that hunk of information in. And you'll recognize this is the topic I showed you before,  where step three and I think, step four are Conref'd in, what that means is if you're going to reuse content a number of times and you don't have Acrilinks or something that will help you manage this, then you put the content that's repeated into a file all by itself that's never itself published. And then you can Conref that content in a million different places and you never have trouble because no one ever deletes your source, and the linking works seamlessly

[00:11:38.860] - Mysti Berry

And I can explain that more later, if I hadn't made that very clear. And then the third solution and the hardest one and the one we had time to do the least often was fix the topic. What are we... Why are we writing about things such that we have to send people hither and yon? Is it an overstuffed topic that really ought to be broken up? Are we, have we sort of forgotten about the customers path through our product and are just relying on our knowledge of things that we know about, not necessarily that the customer needs or wants?

[00:12:16.640] - Mysti Berry

And this is the, I think the place of where I hope we will make more discoveries in the future. All right, so how did the team respond to all this? They showed up, they did the work. I was really amazed at how hard everyone worked to get this thing done. And some of the writers who were working hard actually believe that we were doing the right thing, and they loved deleting those links.

[00:12:43.410] - Mysti Berry

Those links drove them crazy, made their life hard, and they were very happy. But some writers were not so happy. They hated to delete those links. And I think it's the sign of a great technical writer who's worried about every last customer. And if there is, that link will help one customer. They really wanted to do it. But the problem, of course, is if you're spending time on a technique Inline Links that aren't actually ever used by the customers, you're not really saving anyone and you're costing customers your time, I think everybody listening to this call is in a position where they have to choose from the important things to do. They can't do all the important things. 

[00:13:30.310] - Mysti Berry

And in an economy like that and a workflow like that, you can't afford to spend time on techniques that don't pay off. Unfortunately, we didn't know the metrics until after the project and there's been less resistance to letting go of Inline Links since we found out with so few customers actually use them.

[00:13:54.630] - Mysti Berry

And our writers were also sort of feeling in a dilemma because our information model encourages small topics and so a link seems like the right solution. Well, let's talk about this here. But just in case, I'll add a link to this other thing, especially when you're under the gun to deliver before the release goes out. And also, we're highly configurable, many featured product and customers have tasks they need to perform that span across many of our features and writing doc like that is the hardest to do without links. 

[00:14:34.650] - Mysti Berry

And we absolutely did not find all the answers, especially for things like we'll create a single feature and it's pretty isolated, and then suddenly that feature is useful to many other features in the product and it becomes an acquired service. And revising the basic structure of content when that happens is a challenge I don't have the answer for. And also we require a help link on every single Web page in our UI, and that's because we don't want any pages showing up that writers have never seen.

[00:15:07.430] - Mysti Berry

But we know that customers aren't even clicking the help link on those pages. So what's the best solution there? We didn't come up with a good answer for that either. And it was a hectic project that actually surfaced this metric of the links being clicked so seldom, and I also had a writer come up to me in a later conference of VMware writers and she said, you know, I didn't really believe you so we ran some tests and sure enough, we're finding the same thing. 

[00:15:36.410] - Mysti Berry

The higher the link is in the page, the more likely it is to be clicked. But it still clicked way below that 80, 20 cut-off that a lot of us sort of choose our priorities by. There is a movement at our job to implement the DITA feature Chunking, which is where you write in small topics, but they are assembled and output as a big, long HTML page.

[00:15:59.990] - Mysti Berry

And while I think that's perfectly appropriate for this long, complex feature spanning tasks that our customers have to do, the potential for misuse as our own super craziness about linking was misused is there. And some writers are just really frustrated, they don't want to have to think about whether or not it's OK to put a link in their content. And we're still not quite looking for the better structures that will serve our customers better. We're always improving search.

[00:16:34.880] - Mysti Berry

We're always trying to improve our navigation. But when writers... These are quotes from writers who were very distressed originally. And they just, they haven't quite heard the message yet that it looks great to us as writers, but it's not getting used, those links are not getting used. And one of the second coming is an unfortunate side effect is a few topics went out where the writer, just like I, I don't have time to think about this, I'm just going to strip out the links.

[00:17:10.730] - Mysti Berry

And those topics should not have gone out. And we caught those and repaired them. But that does happen, and some of our writers have put the links back in. And this particular topic is often visited and it's very low rated, and I think it's because it's not helping a customer. And there are things that we could have done instead, like we could have pointed them to the UI instead of giving them a link to content. For example, in this topic.

[00:17:45.950] - Mysti Berry

And finally, the comments we get from customers when they're not happy with some of our content. I researched each one of these and found that links would not have helped them, that we had other problems than that. So if you're going to move forward, if I persuaded you that manually created links in content are not so helpful for your customers and you want to get rid of yours because maybe you're going to mobile or maybe you're having trouble breaking up your content, I have some best practices. 

[00:18:21.890] - Mysti Berry

Remember, you're trying to create ones and publish everywhere. So ad hoc, unstructured relationships between topics tend to be a problem. We had some CCMS vendors come in and the amount of cross-linking we are doing between topics actually broke their product. And they had to do fancy things to even demo their product for us, so it's possible to go crazy on the Inline Link thing and you have to do a lot of manual work if you add a new...

[00:18:54.730] - Mysti Berry

Say you're filtering and you add a new platform, a new tablet type comes on the scene, you often have to do a lot of manual reworking to make those links work again. And refactoring content takes incredibly much longer. But we learned you can't Herd a Tiger or a Writer. They are persuaded by metrics and they are persuaded by customer comments and research. So before you start any project like this, check your metrics first, try some experiments, get some patterns.

[00:19:28.450] - Mysti Berry

It only takes one or two bits of reporting for any writer to say, oh, I see, my customers aren't really using that the way I thought they were. And you can't say too many times, look for 30 or 40 years, ever since they invented the Internet, links were the right answer, but now we know they're not always the right answer. So you were doing your job fantastically, but the requirements have changed because some writers heard you're a bad writer instead of we need to get rid of these Inline Links.

[00:19:58.810] - Mysti Berry

And I know the first time I realized I was overusing links, I was very embarrassed. And you have to give writers alternatives. If you're not going to give the customer links, what are you going to give the customer? Just taking things away from the customer will not work for those really good writers. Also, invest in project management. It took 100 percent of my time and I think about 50 percent of a manager's time just to organize the data.

[00:20:24.610] - Mysti Berry

Now, if you have fewer than forty-five hundred or so topics, you may be investing slightly less time, but every... And you probably know this if you're a manager every minute you spend project managing is an hour, you save when the project actually gets underway. And also try working on a small hunk of content to show the writers what it looks like to show yourself that, especially if you can put it out there on the Web to show that your numbers aren't going to fall and they may actually rise or to demonstrate that, yes, this saves us 10 percent of our writing time.

[00:21:01.170] - Mysti Berry

Yay! We couldn't do this because our content was so entangled with Inline Links that we couldn't pull... There was no one hunk we could pull out. And if you're in that situation, then you probably really need to look at getting rid of your Inline Links. And the last thing you can try is create a new doc in a new style without the links and show how it works for the customer. All you ever have to do to cope herd a writer, myself included, is show them how it's better for the customer.

[00:21:30.280] - Mysti Berry

And even if you aren't authoring an XML yet, if your company grows, if you are going to end up translating into many languages, if you are going to go to mobile where you're, what content you're publishing with product, it's going to probably radically change. You want to get rid of inhibition like these Inline Links, or if you're just getting started and you haven't used many Inline Links, keep doing that. Just don't go crazy with the reuse, which is tempting whenever anybody gets XML in their hands on structure FrameMaker.

[00:22:08.320] - Mysti Berry

And that's it, so it's doable and it's...This project and I wasn't thinking that much about Inline Links when the project started and it completely changed how I looked at how we create content and how we make it easier for customers to find it. And I hope that you have a similar experience. 

[00:22:29.770] - Liz Fraley

Wow. Misty, thank you. I knew the numbers were low, but I had no idea they'd be that low. 

[00:22:39.380] - Mysti Berry

Yeah, and if people run numbers and find that they're higher, I want to hear from you because we really want to understand what's going on because, yeah, those are incredibly low numbers. And one writer doing one ad hoc investigation at VMware found similar results.

[00:22:53.800] - Liz Fraley

Fascinating, you know, and I just, it makes me think about reality in a way that I never quite did, you know, think about how do I read through documents and I've read your content and I'm just like, you know, it sounds right even though it feels really weird.

[00:23:09.870] - Mysti Berry

Yeah. We're in a library space when we're writing and our customers are in a I'm going to get this done and get back to my day job space. And they're quite different sometimes.

[00:23:18.490] - Liz Fraley

Absolutely. All right, folks, type your questions in. While you're doing so. It usually takes a minute or two. Let me tell you what's coming up. Our next TC Dojo, we're talking about minimalism and we've got another visiting Dojo Expert Andrea Wenger. She's going to talk about when you have... When the space really matters. She's got to write documents that go in you know, small boxes. What that changes how things are. But it also affects writing for mobile devices and for the web. Same kind of things apply. So that's what's coming up there. 

[00:23:51.360] - Liz Fraley

And just a note about our TPC Affinity Groupss. The next one comes up on October 20th. These are virtual discussion groups. It's where you bring your own problems to the same confidential support environment with visiting experts and fellow members. And you get your questions answered about your life and your experience, your issues, not just what someone else is teaching you. And the last one before we read the questions, TC Camp is coming up in January for everybody in the Bay Area.

[00:24:17.920] - Liz Fraley

And even if you're not, maybe you should be for that weekend, because it's why we hold it on a Saturday. TC Camp is an unconference, very different kind of experience, Mysti is one of our TC Camp volunteers in her oh so copious spare time, and we are grateful for it. It's a practically free event that is unbelievably dynamic and timely topics that happen in the agenda spaces are voted on by the attendees of the daily events. You're guaranteed every session is relevant to you.

[00:24:46.840] - Liz Fraley

All right, thanks for coming. Here's Mysti's contact information, you can always send her questions if we don't get them answered now, but we've got several ready for you already. One is a big, fabulous. Well done. Always like to give those away. All right, so here's one oh, this is a good one, actually. Can you please explain again how you get customers back to the UI?

[00:25:12.140] - Mysti Berry

Yeah, that was a tricky one. So think about assuming you have a Web based product and the customer logs in and they're, say they're adding a new customer into their contact system. And something goes wrong on that page. They click the help link and the help link says, oh, you have to go back a step, go to the setup page and turn this widget. And you can either put a link to the content that tells them how to turn that widget or you can just point them to the location in the UI.

[00:25:49.760] - Mysti Berry

This is go back to the setup page under customers and turn on the widget. So they get back to that page and they're back on task and maybe they're an expert on that page and they don't need any more help. Or maybe they're a beginner and they can click the link on the new page in the place where they're going to do the work. I hope that makes sense. 

[00:26:10.270] - Liz Fraley

We will find out. OK, yes, I get it. Great. 

[00:26:15.160] - Liz Fraley

Next question. Are you familiar with what Mark Baker has been saying in favor of Inline Links? Every page is page one, that kind of thing.

[00:26:22.330] - Mysti Berry

I absolutely adore Mark. I got a chance to meet him for the first time. And I think to a very great extent he is correct, especially since we know that over half... I think I'm allowed to say this, over a half our customers get to their first page in the help from search, not from you are clicking around. But I think there are limits to every page is page one, I think especially in this context of tasks that are complicated or require you to visit different parts of the product.

[00:27:02.210] - Mysti Berry

I think that starts to break down a little bit. But he has done a ton more research. And I actually read his column on LinkedIn and ask him annoying questions all the time. So I think to a very great extent he's correct. 

[00:27:16.100] - Liz Fraley

I've been trying to get him to come in and be a visiting Dojo Expert. Maybe now I can cause everybody wants to hear from him. That's awesome. All right. Somebody wants to see that slide with the metrics again.

[00:27:29.690] - Liz Fraley

I want to make sure everybody knows the video will be posted to our YouTube channel, The TC Dojo. And Mysti has her slides on slide share. I think these have been up for a while. Yes? 

[00:27:39.450] - Mysti Berry

Yeah 

[00:27:40.480] - Liz Fraley

Yeah, OK, so we'll switch back to you and let you put that slide back up and in the mean time...

[00:27:47.210] - Mysti Berry

I'm sorry, which slide did you want? 

[00:27:49.170] - Liz Fraley

The slide with the metrics I think. 

[00:27:51.280] - Mysti Berry

Oh yeah. 

[00:27:51.780] - Liz Fraley

Just let me switch back to that one. We're switching. I will ask the next question. Here we go. it should be up on there. All right. Our next question, any comments about the number of inline links in Wikipedia? 

[00:28:14.050] - Mysti Berry

So let me just find the metric with the full, I mean the page with the full metrics that you quoted. Sorry, I'm making you look at my whole talk again up there we go. So the interesting thing about Wikipedia is when you go to Wikipedia, your task is to read and I think if the task is to read, if the customers task is to read, then you can get away with more links.

[00:28:41.590] - Mysti Berry

And there's nothing else to do in Wikipedia but read and I've got lost happily, joyfully lost in clicking links. Oh, following those paths. But when a customer is working with software and I'm sorry, this type is  super software centric, I think the same thing applies for hardware and middleware. But if the customer's task is get this thing working. They're not really in that read space. They're in that do space. And the mental models and the mental processing is actually different.

[00:29:11.470] - Mysti Berry

And it's kind of hard to go back and forth, which is how we stumbled across the get them back into the UI. Because then they're back in the space that they started with and they're not being forced to switch to read and research mode. That's what we found. 

[00:29:26.620] - Liz Fraley

That sounds about right, it's kind of a different model of behavior, 

[00:29:31.270] - Mysti Berry

Yet you find there's many more links in reference content as well. 

[00:29:36.280] - Liz Fraley

Right, cause you're digging for something. 

[00:29:39.090] - Mysti Berry

Exactly. 

[00:29:39.670] - Liz Fraley

Yeah, cool. Oh, here's a good one. Are related linked tables, just as unused as Inline Links? 

[00:29:48.660] - Mysti Berry

They are far less problematic for restructuring content. And we even reuse, I think, our relationship tables in places, which is kind of esoteric, but they're much safer, and we ended up doing a lot of that when we... I don't know why I didn't mention that, but when we removed the links, sometimes we just put See Also links in. Now when you're trying to use that DITA map in a new context with FAF, you know...

[00:30:24.160] - Mysti Berry

There are places where even the rel table can cause you a little bit of trouble, but it's never as bad as Inline Links. And we do have rel tables everywhere still. 

[00:30:33.420] - Liz Fraley

All right, here's one, what do you use to group topics if you don't have a topic that is all links? 

[00:30:46.300] - Mysti Berry

We do have a table of contents. And we do now use sporadically, we're trying to push it everywhere that DITA feature called I always call it by the way, I just call it parent-child linking so you can turn on a little value in the DITA map.

[00:31:08.200] - Mysti Berry

And it says, OK, at the bottom of this topic for all the children, I'm going to put in a link in the short desk. And so that's automatically generated. The writer never has to go and edit anything if they pull one of those child topics out. And so that's what we were doing for our map topics, because so few people were even clicking the links in those map, not literal map, topics, but those you know, navigation topics that we wanted to do something automatically and not have the writers invest effort in it.

[00:31:38.490] - Liz Fraley

Yeah, that sounds like a great idea. All right, here we go, Misty, I find that even reading text with links is very distracting. A general, generational issue, maybe. Great presentation. Thanks.

[00:31:53.190] - Mysti Berry

Aww, you're welcome. And I've noticed people are starting to make the, this is a link formatting software. So they're getting rid of the underlying they're just changing the color, which can be an accessibility issue. But, yeah, it does slightly distract. I think and I don't even see links now my brain was like either be distracted or don't even see it don't even perceive links, so I don't click for real.

[00:32:18.830] - Liz Fraley

Interesting. I don't know what I'll do, makes me want to think about it, though.

[00:32:25.010] - Liz Fraley

You try to get users back into the UI, but sometimes the UI is broken at SFDC, do you get any input into changing the user experience to reduce the need for help?

[00:32:36.770] - Mysti Berry

We are so lucky at Salesforce. The writers are embedded in the scrum teams. So from the moment the product owner says we're going to make this feature until we deliver it, the writer is there and we do, do that influence for products that have been out or sorry, features that have been out for a while. It can be very hard to drive change. And yeah we're often... I think that's another frustration for writers as they're often in the position of trying to make a not perfect customer experience, better with the doc and I just think there's only so much you can do. And if it's taking right or time or making contact hard to refactor, then you don't really have to look at that. Yeah, it just happens, I mean, all our futures are perfect, but, you know. 

[00:33:29.570] - Liz Fraley

They keep coming in and you got all the answers, that's awesome. Another argument for fewer inline links in PDF output. By default, Adobe Reader defaults do not enable previous view back button on the toolbar. 

[00:33:44.780] - Mysti Berry

Oh, I didn't know that. That's a great point. 

[00:33:47.700] - Liz Fraley

I didn't even know that was there until recently. And apparently it is, you can just when you click through it back, it goes back and like I've been using this for years, I didn't even know they added that. 

[00:33:57.520] - Mysti Berry

I learned something today, too.

[00:33:59.230] - Liz Fraley

That's great, right. All right. Do you ever include mini-talks in topics, or is this what you're referring to when you talk about navigation topics?

[00:34:08.440] - Mysti Berry

We did a lot of hand-crafted mini-talks and we are stepping completely away from that. And if it's a situation where a mini TOC is appropriate, do it with XML and that little widget in the DITA Map that says, yep, print a link in a short desk for all these topics. And we even recently just changed our PDF output so that it would be more clear that those are just that... That is a mini TOC because out of the box, the XML output just made it look like regular content, especially if you print a PDF and you can even tell it's a link and wasn't making a mess. But now it's much better. But yeah, It's the hand coating that was killing us. 

[00:34:50.840] - Liz Fraley

Right. That is always painful. I think we are finally running out of their steam and energy. I don't see any new questions yet. I will make sure that everybody note looks at that last link. Vote for your upcoming session topics. The TC Dojo is all about you. You tell us what you want to learn. Why should we tell you? Right. I don't know what you should learn.

[00:35:13.870] - Liz Fraley

You should tell us and we'll go find the experts to do that. Oh, look, one last question. Oh, these are great, great presentation. I'm feeling the religion and great presentation Mysti, thanks. 

[00:35:25.960] - Mysti Berry

Aww you're most welcome. 

[00:35:27.700] - Liz Fraley

That's awesome. Thank you so much for coming, Mysti I always learn something from you and it's a treat whenever I get to watch you present. 

[00:35:35.200] - Mysti Berry

Oh, thanks, I really appreciate the opportunity to share all this.

[00:35:39.880] - Liz Fraley

Cool. Thanks, everybody, and we'll see you next time.