Wednesday, March 29, 2006

Raising the Bar for Record Managers

Most people that get involved with genealogy today use a family history application called a record manager to store and navigate lineage-linked genealogical information. There are many to choose from. Some of the more common are Personal Ancestral File (PAF), Family Tree Maker, Legacy Family Tree, AncestralQuest, RootsMagic and the Master Genealogist. There are also a range of online record managers like phpGedView, The Next Generation and PedigreeSoft which in some respects are more cumbersome than the traditional desktop applications but offer the advantage of ease of sharing and collaboration.

While these applications are very effective at organizing lineage-linked data, the user experience and complexity is on par with filing your taxes (my apologies to the makers of these products, I know most of them and hope they don’t take offense at this observation).

There have been a couple of interesting advances in the space over the last few years. Notably, the move toward online record managers (phpGedView,) and research guidance (Legacy 6) definitely show promise. So in the spirit of taking genealogy to common people, here is my top ten list of what it would take to raise the bar for record managers.

Top 10 Innovations Needed in Record Managers

  1. Living memory interview – I have personally done usability testing and observed ordinary people taking 30 minutes to an hour to figure out how to enter themselves and their parents into a record manager. When someone starts fresh in a record manager why isn’t one of the options to start a new file from what I know? This option would lead the user through a nice wizard-like living memory interview.

  2. Path to me – Once you browse a few generations it is impossible to tell which path leads back to me. Isn’t there a simple way to add this bit of information to the UI?

  3. Maps, maps, maps – Google Maps, Google Earth, open APIs, need I say more? Maybe not but I will. Users need current and historical maps for research. Seeing a historical map helps me feel connected to my ancestors. Overlaying data on maps is interesting. For example, migration patterns on maps, plotting the events of an ancestors life on a map, showing the overland trail they used to come west, showing the plat map of the town they lived in. Showing everyone with the first name of Deodat living in the US in 1850 census.

  4. Context – Users need lots of context to hold their interest, keep oriented, and to aid in research. Maps (as mentioned above), history (as mentioned in a previous article - record managers really need to integrate with WeRelate.org), historical texture (music of the time, clothes of the time, the price of gas) and anything else you can think of.

  5. Clue Pad/Scratch Pad – This ties back to the need for context. Users need a scratch pad to keep their clues on. The scratch pad needs to let them get back to the clue in context of their pedigree and the source information they were evaluating. It also needs to let them model simple things like: “The wife of John Smith is either Mary Jones or Mary Johnson,” and still keep them in context of the clues that led them to this theory.

  6. Automatic source citations – There has been a lot of dialog on this blog about self-citing Internet sources. Record managers must support this functionality moving forward.

  7. Improved match analysis – I’m not talking about the underlying algorithms (although they are an area of constant improvement). I’m talking about the user interface. How can a novice reliably and consistently make decisions about possible matches in 30 seconds or less? Here are some rough concepts we’ve played with to try and figure this out. They still need lots of work but show how match analysis can be much more than the status quo screens in most record managers.







  8. Interesting Charts – Lines and boxes were cool when dot matrix printers were the rage. Rounded corners had their day as well. Give me a chart that I can put on my wall and my kids, relatives, friends, anyone who walks in the door will notice and want to look at and not mistake it for an electrical diagram. Add historical texture to the charts, themed backgrounds, etc.

  9. Source citation wizard – While I strongly advocate self-citing sources for online content, the reality is we will need to deal with manual source citations for a long time. Let’s build a simple wizard, similar to CitationMachine.net to make it drop-dead simple to cite a source.

  10. Personal Research Assistant/Research guidance – Legacy 6 is moving in the right direction but their feature set works better for advanced genealogists and not ordinary people. Take a look at Grant Skousen’s Family Finder presentation from the BYU Family History Technology Workshop (should be posted in a week or so) to get a feel for how to deliver this for the uninitiated.



I'd be happy to engage in a deeper conversation of how to take record managers to the next level. Add your comments or e-mail me (lawyerdc@ldschurch.org).

30 comments:

Kevin said...

Great Ideas. I agree that they would be very helpful for us novice genealogists.

I don't believe in Microsoft Windows. I am a firm Macintosh user. There is a wonderful genealogy programs written just for Macs called MacFamilyTree. It is from a german company (www.onlymac.de/html/stammbaum4en.html). It has many features not found in the windows programs and its graphics blow everyone else away. I would highly recommend it for all Mac users. It creates GED files that are usable by all the church programs.

Dan Lawyer said...

Kevin,
Thanks for the tip on MacFamilyTree. I'll go take a look.

Roger said...

The idea that you proposed about being able to compare your entries with data from some other person's files is already in existence at GenCircles.com. Unfortunately, only data currently in GenCircles is compared with your own. How great would it be if Family Search would have such a feature -- a great majority of the duplicated ordinance work that the LDS church worries about could be eliminated.

Anonymous said...

Your observation regarding “path to me” is very interesting.
For several years I have tried several programs to see their navigational possibilities.

I think they can be put in two main categories: ‘pedigree navigation’ and ‘tree navigation’.

In ‘pedigree navigation’ programs you have a pedigree starting in a person and can go upwards (parents, grand parents, …). In the other direction you have to choose which of sibling lead to that person.

Examples: Heredis, TMG, Genbox.

Heredis puts a special symbol in the sibling that leads to the principal person.
If you use ‘Roles’ in TMG, you can mark direct lines with different colours. So you can know which path leads to you.
In Genbox it’s possible to mark with only one other colour anybody you want. So, if you mark your two direct lines with that colour, you can know which path leads to you.

In my personal opinion the best are ‘tree navigation’ programs as Kith and Kin and Family Historian.

In Family Historian you can have 5 types of trees: Ancestors, descendants, ancestors and descendants, all relatives and all persons.

You can freely navigate the tree and see at se same time, for example, your parents, their siblings, your cousins, etc. You can even edit the tree.

In Kith and Kin you have also the tree and you can also navigate and edit it.

It seems that people in general prefer 'pedigree approach’ instead of ‘tree approach’.
Am I right?

I would like to know your opinion.

Claudio Cruz

Dan Lawyer said...

Roger,
I like your suggestion about having the ability to compare your genealogical data with data on FamilySearch. In the future I'd love to see a world where you can be working on your family in your favorite family history application and it can show you interesting things like who else is working on similar stuff, what compiled genealogies, digitized records, etc. on FamilySearch match with yours, etc.

Dan Lawyer said...

Claudio,
Thanks for sharing more information on the navigational features of some of these other programs. I'm not familiar with Family Historian or Kith and Kin. I'll check them out.

Although I can see the value of color coding within a record manager, I prefer something totally automated for the basic task of finding the way back to me. Heredis seems more in line with this approach.

As for navigation, I can't speak too authoritatively because I don't have any structured research to back it up. I can tell you what I've seen in my observations. New users often have no preconceived notion of a pedigree, descendency, etc. They often need to be told things like: "You're at the far left in the middle, the fathers are up to the right the mothers are down to the left."

I believe that pedigree navigation is more prevalent in North America because the Church and other large influencers in the field have adopted that approach. Globally there seem to be many cultures that are much more comfortable with different styles of descendency charts.

One of the disadvantages of pedigrees is they tend to discourage research of collateral lines or descendencies.

I've recently been impressed with many of the advantages of drop-line pedigrees also called family trees. Here's a picture of one just to make sure we're all on the same page. I'm not sure how good these are for navigation (I hope to test this) but they seem extremely effective at providing more context and helping users to broaden the scope of their research.

Dan Lawyer said...

I received a great e-mail in response to this article. The sender points out how frustrating it is in current applications not to be able to easily see the relationship between me and someone else in my tree. This is especially hard as you move out to collateral lines.

There are a few applications which attempt to do this. I haven't seen any that I think do this well. Often record managers that provide some level of relationship calculation have a limited range in which they can calculate a relationship and/or they are very cryptic in their representation of the relationship.

Perhaps the list should be expanded to the 11 Innovations Needed in Record Managers. Being able to display the relationship meaningfully for the end user is critical because one of the key things ordinary people are hoping to achieve by learning more about their ancestors is a greater sense of self. Seeing the relationships to others in the family tree and being able to describe the relationship in words helps an individual to have a better understanding of how they fit into the human family.

Dallan Quass said...

Another possible 11th suggestion is the ability to "sync" your local dsektop client with an on-line record manager, where you can see what changes others have made and accept or reject those changes in your desktop repository. This is similar to what software engineers use when a group of distributed engineers collaborate using a shared on-line repository that can be sync'ed with their off-line desktop repositories.

As for maps, check out http://www.WeRelate.org/sources. Do a search for a place (e.g., England) and a keyword (e.g., Vital) and press the "Map" button. You'll see a google map of the top 100 search results. And since the sources database includes all of the items in the FHLC, it makes for a nice way to search the FHLC. This took two afternoons to implement. I need to add some polish, but I think it shows the power you can get when you have a geocoded places database.

Matt Garner said...

Wonderful suggestions, Dan. As a record manager software developer, I guess I'm in the best position to make these things happen. I appreciate the thought and effort involved in enumerating these features, it saves me a lot of time and money on market research. ;) I don't see any reason why all of these features can't be made available, none are unreasonable. Just need some time, I'll get started on suggestion number one next week and just go down the list. :) However, some, like the context suggestion, are going to take some thought to figure out how to create a usable GUI to express the information.

Part of the challenge in writing software for novice genealogists is to avoid making it too easy to just arbitrarily find and import information. I don't want to encourage sloppy genealogy, and have decided to limit such features until I can create a much more effective source doumentation and merge system. I especially like the merge ideas that Dan suggests, merge is one of the trickiest parts, not necessarily technologically, but in presenting the information in a manner where the user can make the most informed decision.

I'll check in to the automatic source citations. I've notice several blogs discussing it. That improvement should be relatively easy and quick to implement.

Either way, I've got my work cut out for me. Let's get going and take record management to the next level!

Anonymous said...

I'm one of those novices who has been frustrated by 3 different programs -- all otherwise excellent. I'm amazed that i seem to be the only person who has children born to unmarried women. If not, why don't the software developers make it easy for me to turn off the marriage "switch." For example, my grandfather was born when his mother was unmarried. She was later married, widowed, married, and widowed again. The programs either try to make one of the subsequent husbands a father, or tell me i haven't indicated a husband for that period of her life. Trust me, if i ever figure out who he was, i'll enter his name! Until then, i'm bugged by a program to do something about it.

Dan Lawyer said...

Dallan,
I love the google map integration with places and the Family History Library Catalog. Makes it easy to see visually what record sets are available, bordering jurisdictions, etc. This is a great step in the right direction.

Dan Lawyer said...

Matt,
Glad to see you're watching the blog. Context is tricky. I received an e-mail over the weekend from a well-known genealogist. They pointed out shrewdly that there is risk that the push for more context may become a "glut of information that may be impossible to digest."

This comment is right on target. Somehow we need to incorporate the context to help the user see the "interconnectedness" of the data without it turning into noise.

ESM said...

Dan,
I totally agree with you that "simplifying citations" would be a wonderful boon for all that "majority of people on this earth [who] want to know more about their ancestors." Alas, "the Devil is in the details"! Do you have suggestions as to how this simplification might be executed?

CitationMachine.net does offer a "drop-dead simple" wizard. However, there is a critical difference between its offering and the material that genealogists work with. CitationMachine.net handles only the simplest of stuff--the things that high school and college students typically use: published sources, interviews, e-mail, and online articles/databases. But good genealogists use *records.* Fitting thousands of types of original sources created by hundreds of cultures worldwide into a basic wizard (or even a few of them) is a radically different issue.

As background, I have at hand three linear feet of e-mailed requests received from genealogists, asking me how to cite some type of source that is not covered by the 318 models in *Evidence!* or by Lackey or CMS or MLA or AP or NARA or PRO or any of the online guides to source citation. When I culled these requests down to the source types most commonly asked about and then created models that also sample the different media in which these original records now appear (manuscripts, microfilm and fiche, CD-ROM, online databases, online images, &c) I ended up with over a thousand models and a 700-page manuscript.

The struggle to reduce all this to a reasonable number of "simple wizards" that could be built into relational databases has delayed for five years the release of the expanded version of *Evidence!* This winter I finally accepted the conclusion I had been fighting: It simply can't be done. Even within census records, a Bureau of Indian Affairs census cannot use the same wizard as a Census Bureau census, because the basic citation elements are substantially different. A record at the courthouse can't be cited with the same wizard as a legislative petition at the state archives, because the identifiers are too different. An immigration record in the U.S. National Archives cannot be cited the same way as an immigration record in the UK National Archives, because the archival cataloging and retrieval data are radically different. &c &c &c!

Yes, if we feel that the only point to a citation is just to say where we got our info so that some Doubting Thomas can go check us, then we could simply cite the website or the FHL film roll number and be done with it. But, of course, all websites are not created equal and all FHL microfilm numbers do not represent the same type or quality of material. If we believe that a source citation should also enable *us* (and those who use our work) to understand what kind of record this citation represents and the strength of the information that came from that record so that we can make valid judgments between conflicting data, then a drop-dead simple citation to a website or a microfilm roll number (whether it be from FHL or wherever), falls far short of what is needed for reliable genealogical research.

Can you (or Matt Garner) turn Pessimistic Lizzie into a New Believer that "automatic source citations" really "should be relatively easy and quick to implement," to quote Mr. Garner? Believe me, I want to believe it--but I just can't see it!

Elizabeth Shown Mills
*Evidence! Citation & Analysis for the Family Historian* &
*QuickSheet! Citing Online Historical Resources Evidence! Style*

jim thompson said...

The best "drop-down" pedigree I ever saw had time (date) as the vertical axis so that the children of a given couple, for example, would have entries stepping down to the right from the first born to the last. Marriages had similar positions. Made my daddy's wait to age 55 to marry easy to see.

Carmen Johnson said...

Two comments...about sources. I would love to be able to post a census record/source to a family. For example, right now I copy and paste the record into each individual. It can be tedious. Sometimes an entire family record comes from one source. It isn't easy to cite the source on each individual.

One thing that I would love to have is the ability to export out of a genealogy program ie Family Tree Maker, Origins, PAF, whatever into Excel so you can compare city/county locations of some families or various other types of patterns. Either that, or have the ability to query the system for all individuals born in Mystic New London Co., CT between 1650-1680. It would seem to me that one of the most important parts of a Record Manager is being able to query the info and get out what you want!

Dan Lawyer said...

Elizabeth,

First of all, thank you, thank you, thank you, for taking the time to read the blog and give such great feedback.

I do in fact have some ideas on how to simplify source citations for ordinary people. The key to simplifying this particular experience is to focus on the needs of the target audience. In this case, the ordinary person (not a researcher, not a genealogist). Observe the target market and try to understand what is important to them concerning source citations.

This particular target market associates source citations with papers that they wrote in college. They did them as needed to get the desired grade. One of their greatest joys upon leaving college (before or after graduating) was the thought that they would never have to cite a source again.

Somehow they determine that they want to try and trace their roots. As they enter information about their ancestors they may or may not think to add a source citation. Perhaps on their own or more likely with a little encouragement from someone who understands the value of citing sources, they try to capture the source information. In record managers of today, this is likely to be a daunting experience. It is a great divider, some will become frustrated and decide that citing sources is not for them. Others will determine some other way to cite the source that is meaningful to them, a small portion will pay the price to learn how to properly enter the source citation.

Given this scenario, I believe it is better to capture a 'weak' source citation from the novice than no citation at all. I would suggest that the measure of success for ordinary people is whether there is sufficient detail, regardless of the format of the citation, to allow another person to track down the original source.

With this philosophy in mind, we mocked up a rough system for capturing source citations and ran it past some novices to see how they reacted. We were very encouraged by the results. In all cases the novice had entered sufficient information to allow another user to get back to the source.

Our user interface was very simple. There were two questions presented to the user with a free-form text box to capture the answer. The questions:

1. Where did you get this information?

2. Where is this information now?

(There were also questions like When was the record made? and What type of record is it?)

Not necessarily pretty but far more effective than anything we've previously observed with novice users.

If I were to build a quick wizard with ordinary people in mind, I would have it cover a limited number of the most common sources for novices (6-10) with a final option for 'Other' that asked the two questions listed above.

Again, not pretty but very likely to help common people enter sufficient source information to allow another user to find the source.

Dan Lawyer said...

Part 2 of Simplifying source citations...


I also believe that simple user experience can be built which meets the needs of genealogists and researchers that want to cite sources.

Here's what I'd do. Build an online wiki-style community for citing sources. Draw the community together and have them contribute their collective knowledge about how to cite every source known to man. The system on the back end would need to capture this information and structure it in a meaningful way and make it accessible through an API. Record managers and other applications would use the content as an expert system for source citations. Requires more work to bring about. Seems very feasible however and also worthwhile. This should probably be done in conjunction with an effort to build a system (distributed or centralized) which catalogs all known sources in the world.

Dan Lawyer said...

Part 3 of simplifying source citations...

Taking a long view of the problem, as n approaches infinity, all records will be available online in a digital format. OK, so what I'm really saying is that there is a clear trend toward making genealogical sources available online in digital form.

The best way to simplify source citations and meet the needs of both the novice and the expert is to build in (and in this case it is relatively simple) a mechanism for automatically citing the source of online genealogical sources.

In this world institutions which make genealogical records available online can include a rich source citation (including, media type, quality, etc., etc.). The application that the novice or expert uses will automatically read that rich citation and put it into the record manager.

Lynn said...

Some of these are already available on ROOTSMAGIC. You can even color-code the direct line (various colors for various lines). And they have easy to use sources, notes, charts, etc. If you haven't looked at this relatively new program, check it out. It's very good!

Dallan Quass said...

I'd be interested in helping build a "wiki-style community for citing sources," especially since a "system which catalogs all known sources in the world" is already a goal of WeRelate.org. The problem is I'm not an expert genealogist, so I'd need someone (ESM?) to tell me what to do. If anyone with genealogical expertise would like to help direct this, I'm interested. (The tough part of being a technologist but not a genealogist is that you can imagine all sorts of cool things, but you're not sure which are worthwhile!)

Anonymous said...

I personally would not mind having a navigational chart looking something like the following.
http://mkgray.com:8000/pix/familytree.png

Dan Lawyer said...

I really like that nav chart

Chuck McKinnon said...

Unfortunately, point 3 isn't as simple as it sounds. You can't call the Google Maps API from a desktop application, and if you're using it for a website it can't be a premium pay-only site.

Yahoo is a little more accommodating, but there's no guarantee they'll approve a developer's intended use.

I expect that the demand will have either Google or Yahoo creating a more commercial-app-friendly API license soon, but for now, it's easier said than done.

Dan Lawyer said...

Chuck,

I'm with you, the market will demand the ability to integrate with maps from the desktop.

Is it possible however to have a hybrid app - embedded web browser that pulls up a web page within the app that has the integration? I'm not familiar enough with the licensing to know if this is an option.

Chuck McKinnon said...

That's an excellent question, and one to which I don't have a good answer. =) I'll find out.

And in any case, I do expect a commercial license to be forthcoming. If Google or Yahoo don't hurry up, Navteq (the data supplier to both firms) will probably do something themselves. They can't be ignorant of the demand.

Logan Allred said...

Great ideas, though I expect most software developers have at least half of these on their wish list already. Genealogy is complex and the users have such varied goals and experience levels. It really is like filing your taxes!

There are so many things I'd love to do with the record manager I'm just finishing up right now, but I'm hampered from the cool stuff by all the plumbing. I shudder to imagine how many GEDCOM parsers are out there, and probably even more data models. Sadly, a good robust parser with error recovery and data cleansing is still a feature instead of an expected commodity. If there were a more standardized infrastructure, I think many developers could spend more time on cool stuff and less on plumbing.

Also, good UI is hard. Unless the core developers are naturally gifted (I'm certainly not), it's going to take resources, and most genealogy software shops seem to be very small with limited resources.

Many of these could be implemented as modules or plug-ins, but outside of reporting and a little bit with PAF for Windows, I don't really see much plug-in support or modularity out there.

I sure hope the new Family Tree will have an API that the record managers can interact with.

Dilworth said...

I am a Legacy user and I really like the way they address the "path to me" concept. In most of the list type screens all of my direct ancestors are in bold. When I am in the family screen and I want to walk down the path towards me I just click in the blank space to the left of the children. It moves the child that is the direct relative to me up to the parent. To enable this behavior I made a one time designation of my entry as the direct line. This has also helped me find problems. One time I was navigating back and got caught in a loop. After a little research I discovered that one ancestor was listed as the son and the father of the same individual (I had pulled some data directly out of Ancestral File and had not checked this part out yet).

I have also turned on the relationship indicator and so if I look on the family page, just above each of the parents, I will see how that person is related to me. This really does help make things personal. When I am looking at a person and see that she is my seventh great grandmother, I just feel closer to her seeing it phrased that way. When I first turned it on it told me that my wife was my ninth cousin. I had not known that before. With a few additional steps it showed me the chain of relationships to our closest common ancestor.

Jacqueline du Plessis said...

I love the ideas of comparing data in a visual way - once you can SEE information, analysis is so much easier. LOVE the ideas!

Jacqueline du Plessis said...
This comment has been removed by a blog administrator.
Anonymous said...

You know, I've found it interesting that "PAF for Palm" had a simple method to track back to your "home person," but PAF (the 'full' version) did not. If it's easy enough to put in the Palm version, why does the full version give you the option to either 'jump to home' or 'get lost in the tree'?

- Stephen