Modern Software Experience


Family Tree Builder

Family Tree Builder was introduced in 2006 July. A few months ago, Family Tree Builder 2.0 was released, on 2007 Aug 28 to be precise. The most important new feature of version 2.0, the one that the MyHeritage announcement goes on about is Smart Matching.


Family Tree Builder is a free download of less than 20 MB. You do not need to register to download it, but MyHeritage does expect you to register to use it.


Family Tree Builder (FTB) is a multilingual program. The first thing the installation program asks you to do is pick a language from a drop-down list. Now, Dutch is in that drop-down list, but I started by choosing English.
The Setup Wizard picks a reasonable default directory, but you can change it to another. The setup proceeds without offering any other options. It installs a desktop icon, whether you like that or not. Once the Setup Wizard is done, it offers to start the program.

Starting Up


The first time you start up FTB, the program throws up a product registration dialog box. MyHeritage requires registration. Existing members can sign in with their email address and password. As I had already registered with MyHeritage to search their collection of family trees, that is what I did. As soon as I clicked the "Sign In" button, my firewall warned me that FTB was trying to connect to the Internet. I allowed this, and got the FTB Welcome dialog. I then clicked the Next button and was presented with the Quick-Start Wizard.

The Quick Start Wizard offers a choice between importing a GEDCOM, starting a new project, loading an existing project, or loading the sample file. I decided to start with the sample file. The program then loaded the file and, quite unusual and a bit annoying, upon completing the load, presented some file statistics. Really now, I’ll select that dialog box from the menu when I want to know. The sample is a 95-person tree for American president John Fitzgerald Kennedy.

user interface

The user interface seems fairly normal. The main window is divided into two panes, with two different lists to choose people from on the left side and a view of your data in the rest.

The default view shows three generations; two big boxes for husband wife in the centre, smaller boxes for their parents above them and their children below them. This view already shows a lot of information, but there is a slider above it, to show one or two generations more - and That’s one or two generations in both directions, ancestors an descendants. What MyHeritage calls the "Family Tree View" (and, confusingly, sometimes the "tree view") view is really an hourglass display of three, five or seven generations. It is not an all-in-one diagram; it does not show everyone, but only ancestors and descendants of the selected couple, and it does not show all generations, but at most seven.
I quite like this view, but I would like an option to show the other spouses of the selected couple, and the children they had with these as well.

individual lists

There are two lists of individuals on the left side, in what MyHeritage calls the the Selection View. It is actually a selection pane with two views in it. One is a simple list of individuals that you can sort by ID, first name and last name. The other is not really a list, but a tree view of families. This sounds and even looks a bit confusing at first, but think of it as an a navigatable and clickable descendants tree on the left, which you can click on to choose a person for display in the main hourglass view on the right.
If you have a small screen or just don’t like the pane and the two lists it offers, you can minimise it to nothing but two tabs on the left side of the main view to bring the views back.

user interface flaw

That TreeView masks a user interface flaw by compensating for it. You’d expect to be able to click a person in the family view to move it to the centre position, perhaps right-click it and then pick a menu item, but in FTB you have right-click, select a submenu, and then click a menu item. Some may come to like the tree view on the left, but I wondered about the memory footprint of this user interface item. TreeViews can be handy, but they don’t come cheap. You sure don’t need it for the added navigatability, as the central boxes for the husband and wife already contain a clickable list of all their partners, siblings and half-siblings. You only need it because the ability to move one of the other boxes to the centre position is tucked away too deep.

database reload

Moreover, the first time try to use this TreeView, it loads your database a second time, and even does this somewhat slower than the normal load. The immediately obvious problem is that the program is making you wait when you are ready for action, and wanted to select an item, not study that progress bar again. It also takes some additional memory (with the 100k INDI database, usage climbs from 369 MB to 400 MB, That’s 31 MB more), but the real question is why it is loading all that data a second time at all, when all of it is in memory already. That progress dialog box that FTB throws up when you access the TreeView for the first time reflects bad design, perhaps even lack of communication between team members.


And if consuming more memory and making you wait is not bad enough, navigating the TreeView for sizeable databases is hardly possible anyway, because it takes it sweet time to response to even the simplest operation, such as trying to scroll the TreeView to find someone.

more views

There are two other views on the View menu. One is a media view. MyHeritage calls it the Photos and other media view;, and I think they should shorten that to just media view. As the name suggests, it provides an overview of the media in your project, but I have no images in my test file. Most programs are still so bad at handling GEDCOM, have such awkward user interfaces, and so much trouble large databases that I have more than enough to write about already.

Smart Matching

Pearl Street Software

The third, and most interesting view is the new Smart Matches view.
On 2007 August 27 MyHeritage merged with Pearl Street Software. Well, that is the official, nice of way of saying that MyHeritage bought Pearl Street Software, which had put itself up for sale. Pearl Street Software made the Family Tree Legends software and the GenCircles web site, which is known for having pretty good Smart Matching technology. MyHeritage believes they improved it further.

In the Family Tree Builder 2.0 announcement on the MyHeritage blog, Gilad Japhet writes One of the many fruits of this merger has been the addition of Smart Matching™ technology to all MyHeritage products, including Family Tree Builder, and significant improvements of Smart Matching™ added by MyHeritage to a level available no where else on the Internet..

MyHeritage claims that they added Smart Matching to FTB, but the truth is that Family Tree Builder only works with the Smart Matching feature on the MyHeritage site.

individual match

You need to upload your data to perform Smart Matching on all individuals in our tree, but you do not need to upload your entire database to perform Smart Matches on a single individual. At least, that is what MyHeritage suggests. According to the same My Heritage blog post, ..any time you add or edit a person in your family tree, Family Tree Builder will connect this individual or any of its relatives to approx. 180 million people in other family trees published by other users to MyHeritage. This will happen in real-time without slowing you down., and then goes on to say that should upload your entire database to get matches for all individuals.

easy match test

To test the Smart Matching, I navigated to an individual in the database that is listed on the MyHeritage web site, and started to edit the record. I merely added a death date and then saved the record. As soon as I saved it, FTB connected to the Internet. The cursor became an hourglass, and Windows added the dreaded (not responding) phrase to FTB’s title bar, so the in real-time without slowing you down. claim seem no more than wishful thinking to me. The fact of the matter is that I had to wait about four minutes to get results. I am guessing the experience is a lot smoother if you upload your entire database first, but there are other issues with that.

not so smart mismatches

I got three ostensible matches, but not one of these seemed particularly smart to me. Although the name and birth date for the individual I had just edited matched exactly, these three matches where not for that individual, but for three other individuals. For all three, the program noted how many steps they are from the person with ID 1, not how they relate to the record I just edited.

I examined these three matches anyway. I double-clicked the first one to view the smart matches, FTB connected to the Internet again, but all that happened was that the person disappeared from my Smart Match list. I right-click to one of the remaining two and choose "Display in Family Tree". That shows the individual in the centre of the main hourglass view, but FTB does not display any matches there, and when I go back the list of matches, he’s gone from the list. One individuals remains. When I select "Display smart matches" and allow the Internet traffic, a text above the list of individuals reads There are no smart matches for Berendjen Lammers. I click the "Update Now" button to make sure, but it does not help.


I try updating that same person again. I removed the death details I had added and saved again. FTB immediately connected to the Internet and became unresponsive again. After several minutes, it presented the same three matches again.


There are three tabs, called Overview, Matches and Compare. The Overview tabs lists all three person, the Matches tab should probably list all the matches for a single person, and the Compare tabs is supposed to show your tree and theirs, but it does not. Even after selecting an individual from the list and hitting the Update button again, no matches are shown. I do have two individuals with the same name in the tree, but these do not match with other.

One of the names in the list of three ostensible matches occurs just once in my database. When I search the website for the ostensible matching individual, no match is found. When I search for another of the three names, no match is found. When I search the website, I find one match for "Berendjen Lammers", the least unique name of the three, but it is only the name that matches; the birth date, birth place and parents are all different. All in all, FTB’s Smart Matching failed to impress.

the truth about Smart Matching

Perhaps MyHeritage does indeed have the best Smart Matching in the known Universe, but to really take advantage of Smart Matching, you really must upload your data to the MyHeritage web site. MyHeritage claims that they added Smart Matching to FTB, but the truth is that Family Tree Builder only works with the Smart Matching feature on the MyHeritage site.

… many researchers … would like to use Smart Matching to help their research forward, not to publish it prematurely.


If you’ve ever visited GenCircles, you know that Smart Matching works pretty well. As you browse through a genealogy, you’ll often see a Smart Match that links to another genealogy, and any time I followed such a link, it all seemed to match perfectly.
Obtaining Smart Matches for your entire database in exchange for uploading it is an interesting offer, but there are a few issues you should consider before uploading your data.

For starters, it is not just uploading your data, it is publishing your data on their site, for all the world to see. That default behaviour does not make much sense, as many researchers do not want to publish their work until they consider it sufficiently finished, and they would like to use Smart Matching to help their research forward, not to publish it prematurely. A match may lead to a contact and some data exchange, and the other researcher may decide to publish the overlap you researched together, but that is not the same as having your research published when you are not ready to publish it yet.
Moreover, MyHeritage does not post Terms of Service at all, and deleting your tree from is not as easy as creating it; there is no delete button, wants you to mail them instead.

FTB prompts you to upload your entire database, so I guess that includes not just all my mistakes, which others might provide corrections for, but all my private notes as well, which are private for a reason. Worst of all, it includes living individuals, and publicly posting that information, or even just providing it to a third party like, without their permission is illegal.

I am not a lawyer, but it would not surprise me if FTB’s constant prompting to upload all my data is illegal too, not so much because it is a pushy marketing tactic to obtain my data in an effort to expand their site, but because they are encouraging me to perform an illegal act. The very least MyHeritage should do to comply with privacy laws is install a Private and Living filter on the database upload.

works at all?

I may try to take advantage of Smart Matching by uploading some selected fragments later, but for now I simply conclude that FTB does not really support Smart Matching, and that its integration with Smart Matching on the web site is problematic in several ways. I even cannot help but wonder whether FTB’s Smart Matching works at all. After all, FTB failed the Easy Match test, a match even Dumb Matching technology would have succeeded at, so they sure messed up somehow.

All in all, MyHeritage’s promotion of Family Tree Builder, including its Smart Matching feature, as a free product seems less than totally honest advertising to me.

not free

Perhaps it is only the matching of a single individual That’s defective, and Smart Matching works fine when you upload your entire database, but you will not only have to give them your data, you will have to give them your money too, because uploading you data to is publishing your database on, and that is not free. Thus, in practice, Family Tree Builder is free, but its Smart Matching support through the MyHeritage website is not.

Sure, the MyHeritage web site offers free family sites, but only if your database still contains less than thousand individuals, because you just started your research. Once you have a thousand individuals you must switch from the Basic Plan to the Silver Plan, and as soon as you reach 2500 individuals, you must switch to the Gold Plan. That is the lowest plan to support an unlimited number of individuals. It costs several dollars a month. Now, paying for one month to get Smart Matching for your entire database seems totally worth it, but the minimum subscription period is a year.

All in all, MyHeritage’s promotion of Family Tree Builder, including its Smart Matching feature, as a free product seems less than totally honest advertising to me. It is closer to the truth to say that you get a limited free edition, but will have to upgrade to the paid edition to take advantage of its best feature. It is a very nice model, as you only be paying for that additional feature when you want it enough to pay for it.

Currently, the spectacular failure of individual matching does not convince that FTB’s Smart Matching works at all, but MyHeritage will probably fix whatever they did wrong in the next upgrade. Once the free individual matching works fine, and thus convinces that MyHeritage has solved its initial implementation issues, I’d be willing to pay some money for Smart Matching of my entire database against the databases on their website, especially when I can sort all the matching individuals by how related they are to me. I’d just like to do that without having to give them my data or publish on their site.

Smart Matching Service

I think MyHeritage should consider offering a Smart Matching Service separate from publication of data on their website; upload my data, don’t show it to anyone, perform smart matches, download these to Family Tree Builder, and delete my data from their servers again. Once they have it working for Family Tree Builder, they may want to open up their API for integration into other software products, and thus gain wider exposure for their paid service, which in turn leads to wider exposure for their web site.
I am well aware that I am ignoring a myriad of little details, but the basic idea is solid, and solving those details is MyHeritage’s job, and what you’d ultimately pay them for.

Alas, current reality is that they still need to get it to work reliably. Claiming three matches for persons I did not edit, but failing to show any match while not even discovering the Easy Match for the person I just edited is not Smart Matching, but less than Dumb Matching.


22 languages, no English

Family Tree Builder has a button bar, and to the right of the button bar is a drop-down menu that lets you select a language. The American instead of English flag next to "English" in the drop-down list confirms that FTB does not support English, but Amglish only.
The 23 languages FTB claims to support are "English" (Amglish), Hebrew, French, German, Italian, Spanish, Portuguese (Portugal), Russian, Dutch, Greek, Turkish, Yiddish, Portuguese (Brazil), Norwegian, Danish, Lithuanian, Polish, Swedish, Ukrainian, Hungarian, Czech, Finnish and Afrikaans - and That’s the order in which they appear on the drop-down menu.

Note that Portuguese is listed twice. Yiddish and Hebrew are two different languages, but Portuguese (Portugal) and Portuguese (Brazil) are merely two different locales for one language. Those 23 choices are not 23 languages, but 23 locales for 22 languages. English is not supported, Amglish is mislabelled as English, and whether the Norwegian is Bokmål or Nynorsk is not made clear.


You can switch between languages without restarting the program. The flags in the drop-down box are no luxury, because even that list of languages in shown in the chosen language. You may need those flags to make sense of that menu after selecting a language or even script you don’t know. When you select Hebrew or Yiddish, the entire user interface becomes right-to-left instead of left-to-right. For some reason, perhaps the TreeView, the program reloaded the database, and displayed a progress bar that filled up from right to left.

I had a brief look at the Dutch interface and it seems okay to me.


When I select the first item on the Help menu to see whether the help file is in Dutch too, FTB tells me that the help is on the MyHeritage web site, and offers to download it. It’s good that FTB offers to download it, so that the help becomes local, but it is no good that it is separate download. That means means you can’t get to it if you happen to select this on a laptop without Internet access - unless you’ve downloaded it already. The help file should be included with the installation package.

The help is Amglish only. There is no Dutch help, there is not even an English help.
In fact, there is no help file at all. The download is not a Windows help file, but an user manual in Adobe PDF format. A user manual is handy to have, but it is not the same thing as a help file. If I had not relied on the first menu item of the Help menu item to bring up the help file, but actually read what it says there, I would have know this before downloading the file. It does not say "Index" or "Topics" but "User’s Guide...".
Weirdly, even after downloading, FTB keeps trying connect to the Internet before loading the local copy.

GEDCOM import

I tested GEDCOM import with two files I’ve mentioned and used in earlier reviews. One is a file of about 1 MB, the other contains about 100.000 individuals. I’ll refer to these as the 1 MB GEDCOM and the 100k INDI GEDCOM.

memory usage

I had already noticed that, upon loading the sample file with less than 100 individuals in it, the Task Manager showed a memory usage of roughly 41 MB that reduced to 31 MB when I closed the file. FTB seemed to need almost 10 MB for a tiny tree of less than 100 individuals. That suggests a memory usage of more than 100 KB per individual, so I kept an eye on the memory usage.
Memory usage did not get that bad again, so my first guess about what’s going on here would be that FTB’s memory usage is fine in general, but perhaps not so good when it has to deal with files that include images.

data in program directory

I noticed another thing before doing the actual import test. FTB asks for a project name and then makes a file - in a subdirectory of its program directory. You can change the so-called Project directory in the Options dialog box, but that does not change the fact that the default directory is wrong. Data should not be stored in the program directory.


Import of the 1 MB GEDCOM file into a new family file takes 14 seconds - That’s including the delays caused by multiple dialogs that popped up during import. FTB shows GEDCOM statistics when it is done, as well as a "loading time" of 2,3 seconds. That’s nonsense. Perhaps FTB will get close to that when MyHeritage take the annoying superfluous dialogs out, but the actual time it takes to import this file is 14 seconds.

FTB first imports the GEDCOM file and then saves the project. I later discovered that most of the "saving" time, FTB is not really saving at all, but preparing to rewrite the entire database, building the file it is going to write in memory, and only when the progress bar is around 99 % does FTB actually write that file.

setting up

You can create a new family project and then select a file to import. Getting this far already takes a few dialog boxes too many; but let’s concentrate on the import experience itself. Once you’ve set the import up, FTB shows the file you’ve selected for import in a dialog box, and you can choose "Next" to proceed.

empty project

I prefer to perform GEDCOM import speed tests by importing into an empty existing project, as that avoid the time it takes to create the project, and the dialog boxes associated with it. Although FTB lets you select a GEDCOM file to import into an existing project, once you choose that Next button, FTB pops up a messagebox saying "There is a genealogy Project currently open. In order to proceed, it should first be closed." and continues, as if I indicated that wish, "Are you sure you want to close project "Test1MB" ?". Apparently, FTB is unable to import a GEDCOM file into an empty project, so you choose "Yes".

FTB does not ask immediately ask for a new project name. It first throws up a dialog showing the information in the GEDCOM header. You need click "Next" to proceed. The next dialog box is definitely unusual; it asks you to choose the primary language used in the GEDCOM file. The default is Amglish (mislabelled "English").

FTB says it cannot

When you choose "Next", FTB ask for a project name, and provides a default based on the GEDCOM file name. I have to wonder why they want to make it so complex. Suggesting a name for my project is nice when I import to a new file, but I had already created an empty project to import to. Some programs cannot add GEDCOM data to an existing database, and there are reasons why you generally shouldn’t do it anyway, but refusing to import into an empty file the program itself just created, and then demand that I create another empty file to import into is ridiculous. Apparently, and this is where it gets truly confusing, the makers of FTB agree with me. Although this dialog box series started by demanding that I close the empty project, the dialog box that asks for a new project name has an option to import into the still open project. I just need to choose "Merge into project "Test1MB"".

This is needless complexity. If FTB can import the GEDCOM file, why do I get the run-around through all these dialog boxes? Why does it not just do it and why does it start by telling me it cannot do it?

After working with the program for a while, I now think that it all it really wants to do is save the file, and that it so does so by using part of its file-closing logic, because FTB always saves files upon close. The only real failure seems to be insufficient distinction between saving and closing. If FTB acted just the same, but with a different (save instead of close) message, it would make more sense already.

Convert to Notes

There is another option on that dialog worth mentioning. Several genealogy program disrespect your database so much that they add comments to your records for anything they cannot handle, which often turns out to be a surprising amount, and you end up with useless comments all over your reports.
FTB isn’t quite that bad. It offers you a choice between "Convert to Notes (recommended)" and "Discard". Sadly, FTB makes "Convert to Notes" the default choice. That is so wrong. The choices should be "Convert to notes (NOT recommended)" and "Notes in import log (recommended)". Actually, converting import problems to notes should not even be an option. The program should recognise all standard tags and all unrecognised tags should always be listed in the import log.

character encoding

There is another option on that dialog that should not be there. It allows you to specify the character encoding of the file you import. Luckily, the default is auto-detect. When you look through the list of option, you’ll notice that FTB supports conversion from various code pages. That seems nice, and it makes some sense to ask for a code page when the GEDCOM file lists the character encoding as "MSDOS" or "Windows" but when it lists "UTF-8" or "ANSEL", there is nothing to ask, and the program should just import it.

Moreover, asking for a primary language in one dialog box and then asking for a code page in another dialog box is not exactly the pinnacle of logical and user-friendly design.
language and code page

Moreover, asking for a primary language in one dialog box and then asking for a code page in another dialog box is not exactly the pinnacle of logical and user-friendly design. If the program lets the user choose both, both choices should be offered on the same dialog, the program should help you select the right language / code page combination, and warn you when you select an unusual or impossible combination, such as Arabic on a Russian code page.


Only after you have suffered through all these dialogs and click Next again does FTB actually import the file. you’ve probably spent about ten seconds from the moment that FTB started processing the file by reading the header till this moment, by making choices, correcting erroneous defaults and clicking Next. When you do click Next once again, it shows a progress bar for about a second or three, and then shows a dialog with GEDCOM file statistics and the claim that the import took 2,3 seconds.

actual time

There is another load of dialog boxes to wade through once the import has completed. Although some of these are about essential import features, such as saving the import log, I have not counted the time it takes to get through these, but I do still measure the import time from clicking the Next button after which it starts reading the file to display the GEDCOM header till completion of the import. So the actual import time for the 1 MB GEDCOM is 14 seconds.
If MyHeritage bothers to clean up their convoluted maze of import dialog boxes, that time will come down significantly.

import log

The dialog box with the statistics and the dishonest import time sports a "View Issues" button. When you click it, the program does not show an import log, but another dialog box with a TreeView of the issues it found. At this point, there is no import log yet! Only if you decide to click the "Save Report..." button does FTB write an import log. Strangely, it does not default to the same directory as GEDCOM file you imported, but not to the same name, it expect you to pick a name. I should not need to pick a name, nor should I need to confirm that I want an import log. The program should always write an import log with the same base name as the GEDCOM file, just a different file extension. If FTB crashed before reaching this point where it begrudgingly allows you to save errors and warnings to the import log, there would be no import log at all to give any hint as to how far it got or what problems it encountered.


Now, in the case of this file, the GEDCOM import rightly complains about some badly structured dates. However, the log file FTB writes is just a dump of the TreeView it created, and is therefore just as uninformative as that TreeView. It does not show either the line number, the line or the record the date appeared in. It merely notes that it encountered some bad dates and then presents a list of the dates it could not handle. Figuring out just where in the GEDCOM file these dates appear is left as an exercise for the user. The error message is crystal clear, but within a line number, a line or a record, the import log is close to useless.

You must go through this dialog box and must pick a file name by hand, or you won’t have an import log at all, and you then need to click OK, to get back to the dialog box with the statistics, before you can close that one.

dialog box parade

The dialog box with statistics has a "Finish" button, but don’t get your hopes up, it is still not the end of the dialog box parade. FTB pops up a "Publishing Wizard" that prompts you publish the data to a family site - on of course. This has nothing to do with importing data, and therefore sure to annoy and alienate users. Only when you dismiss that pushy dialog box is the import truly done. Annoyingly, FTB pushes another publishing dialog in your face when you close the project without publishing it.


The progress bar for the 100k INDI GEDCOM shows that FTB’s GEDCOM import is faster than most of the competition. It seems to move at roughly one percent per second. And indeed, FTB claims an import time of 100,8 seconds. I tried to handle all the dialogs quickly, but I still measured an import time of 107 seconds. I haven’t done a measurement recently, but I seem to recall that PAF takes about three or four minutes to import its own GEDCOM, and would make FTB about twice as fast.

The measured import speed is about 935 INDI per second. If MyHeritage cleaned up their import dialog mess and reduced the claimed import time to 100,8 second, that would become about 993 INDI per second. That is still less than 1000 INDI per second, but close.

import errors

The import log lists some unstructured dates, a few unrecognised tags and some miscellaneous errors. The complaints about the unstructured dates are mostly about dual dates such as 20/22 Sep 1972, and the unrecognised tags are all PAF-specific GEDCOM extensions. Weirdly, although it does not report these for any other issue, FTB does report line numbers for the unrecognised tag. The miscellaneous errors are the most interesting ones.

missing mutual references

FTB reports four missing mutual references. It is possible for PAF to get messed up and then produce a messed up GEDCOM file, in one of the two links between a family and an individual is missing. The only other program I know that discovered and reported this problem is Behold. Most programs do not seem to notice.

Invalid ANSEL sequence

FTB goes on to report two issues no other program ever reported. It reports Invalid ANSEL sequence (226.32). In the few ANSEL tests I have done with a variety of programs, most programs make no effort to check the encoding, but blindly assume that the all values make sense, and then produce, cough, varied results. FTB is quite right to complain about errors in the encoding of the GEDCOM file, the problem with its complaints is that there is no error.

Decimal 226 is hexadecimal E2. ANSEL code E2 is the Combining Acute Accent. Decimal 32 is hexadecimal 20. ANSEL code 20 is ASCII code 20, the Space character. Thus, (226.32) is ANSEL sequence E2 20, a Combining Acute Accent on top of a Space character. FTB claims that it is an invalid ANSEL sequence, but FTB is wrong. It is a perfectly valid sequence, that correctly encodes the Acute Accent (U+00B4). The real problem is that FTB fails to recognise a correct sequence. That is a limitation of FTB, not an error in the GEDCOM file.

If FTB does not recognise the correct encoding for Acute Accent upon import, you cannot help but wonder how it exports an Acute Accent. The answer is that FTB lacks support for export to ANSEL.

Undisplayable Unicode Character

FTB also has several complaints about undisplayable Unicode characters. In actual fact, few Unicode characters do not display, but for those that do not, their lack of display is a Good Thing, not something to complain about, so that is probably not what FTB means.

One specific error FTB produces is Undisplayable Unicode Character: 264 (227,67).
Decimal 227 is hexadecimal E3. ANSEL code E3 is the Combining Circumflex Accent. Decimal 67 is hexadecimal 43. ANSEL code 43 is ASCII code 43, the Latin Capital Letter C. Thus, (227,67); is ANSEL sequence E3 43, a Combining Circumflex Accent over a Latin Capital Letter C. Decimal 264 is hexadecimal 108, Latin Capital Letter C with Circumflex, and that E3 43 sequence is the correct ANSEL encoding of Unicode character U+0108, Latin Capital Letter C with Circumflex.

There is nothing wrong with the GEDCOM file, so there is no error, nor does FTB report an error, FTB reports a warning. However, that warning is wrong. Unicode character U+0108, Latin Capital Letter C with Circumflex is perfectly displayable.

This message actually reveals a serious design flaw in FTB. There are several similar messages in the import log. What these have in common is that the Unicode code point for the character needs more than two hexadecimal digits, because it is not one of the first 256 (decimal) characters. What these mistaken messages reveal is that FTB, although it has international menus, and runs on a Unicode operating system, is not a Unicode program, but a code page-based program. It supports nothing but the 256 characters of Windows code page 1252, which is roughly the first 256 characters of Unicode.

The message FTB produces is wrong. These Unicode characters are not undisplayable, they are unrepresentable in the limited Windows 1252 character set that FTB still uses. FTB should issue a warning, and should make it clear that this is a fundamental FTB limitation, not make it sound like there is something wrong with the characters.


The 100k GEDCOM import revealed a silly defect in the import. The two persons shown in the centre of the screen do not have ID 1 and 2, but ID 65.537 and 65.538, that is exactly 65.536 more. This suggests that the program is using both 16-bit and 32-bit numbers for the ID, instead of 32-bit numbers exclusively, and such mixing of 16-bit and 32-bit arithmetic makes me fear that the program cannot be relied upon to make the right links once a database contains 65.536 individuals or more.


I set the display to seven generation and navigated to some big families with a lot of children and yet more grandchildren. The screen got a bit busy and the resulting graphical display was really too wide for the limited viewport. FTB was no speed demon with just few hundred individuals on screen, but FTB displayed everything quickly enough and scrolling worked just fine.

GEDCOM export

Remarkably, GEDCOM export takes just as long as GEDCOM import or saving a family file. FTB offers several options during GEDCOM export, including the ability to private the file.

When you export to GEDCOM, FTB prompts you for submitter information and remembers it for the next time you export. That’s good. It also prompts you for a photo directory, even if you have no photos in your file. That’s bad.

The resulting GEDCOM file looks good. There are a few proprietary extensions, but they are few and start with an underscore as they should. FTB uses the _UID extension like PAF and Ancestral Quest, but, unlike PAF and AQ, actually formats the GUID as a GUID. That difference in formatting makes it incompatible with each other, but FTB quite correctly did not complain about that when reading the PAF GEDCOM.

The GEDCOM does not support either ANSEL or UTF-8. FTB does not even ask what character encoding you’d like to use, but always writes "ANSI". This is a major limitation. You should not use a genealogy program that cannot export to ANSEL or UTF-8.


FTB is not Unicode based. FTB is a code-page based design that uses Windows code page 1252 ( "Windows ANSI") internally.

The FTB user-interface is Unicode-based, but the FTB internals are not.


Although I cannot give a thumbs up to any code-paged based genealogy program, because it cannot fully support ANSEL or UTF-8, I can still understand such a design decision in a program That’s already 15 years old, or one based on database platform that doesn’t support Unicode, but the facts are that FTB is just a few years old, and not build on top of an aging database platform either.

What makes this self-limiting design decision more puzzling is that FTB has been created with tools that made it easy for them to create a multilingual user-interface. The FTB user-interface is Unicode-based, but the FTB internals are not.
What makes it even more puzzling is that MyHeritage is a Israelian company; Windows ANSI may seem good enough to many Westerners, but it is immensely limiting to anyone who grew up with the Hebrew alphabet.

no Hebrew

The facts are obvious, but it is so illogical, that I still decided to have a look at what happens when I try to use Hebrew. Wondering whether FTB does perhaps support a Hebrew code page, I decided to do a test while using the Hebrew user interface .I recognise Hebrew, but I do not read or write it, and am not used to reading from right to left either. I have to rely on my understanding of the program so far, but the test is simple. I copy some Hebrew text from the web and paste it into NotePad. Like Firefox, NotePad displays it correctly, thus confirming that I have the necessary fonts. When I paste the same text in Family Tree Builder it becomes a bunch of questions marks with some spaces in between. When I examine the database to make sure, I find that FTB really saved but a bunch of question marks (ASCII code 3F). Even when you set the database’s primary language to Hebrew and use the Hebrew user interface to export the data, FTB still writes its GEDCOM files in Windows ANSI.

Windows maintains a default code page for non-Unicode program and I can set this to Hebrew to simulate a Hebrew Windows, but it still wouldn’t work. The database FTB uses is a Windows ANSI database, an the GEDCOM files that FTB writes are "ANSI" GEDCOM files; even if I could get Hebrew in, I would not be able to get it out again.

I cannot describe the creation of a fairly new code-page based program with all the limitations that entails as anything but a fatal design flaw. It makes the program unsuitable for most people, including the creators themselves.


Family Tree Builder does not lack reports. It can produce Family Group Sheet, a Relationship Report, Ancestors report, Descendants reports, a Timeline and a Descendants Summary. All reports are shown in the main window.


You can save reports to three formats; RTF, PDF and HTML and FTB offers to open the file after saving it. I tried all three, and even the tabular format of the Family Group Sheet was rendered just fine -with one exception; in the Family Group Sheet the footnotes were not shown as footnotes, but inline, as run-on text in the HTML output, and that does not look good.

Family Group Sheet

The Family Group Sheet has a clean look about it, and default to including a picture. Even a report as simple as this has a bunch of options to tweak it to your liking. You can omit the picture, change the fonts, optionally include notes, opt to include sources as endnotes or footnotes, and a few more.

Relationship Report

When I tried the Relationship Report, the progress dialog box that FTB put up it quickly made apparent that it would take more than an hour to generate it. I decided to do some necessary shopping, and when I came back and looked at it again, it was still running. I did not time it exactly, but I can tell you that it took more than two hours, and that when it was done, Task Manager showed FTB to be using 758 MB of RAM.

If I had not recently upgraded the PC from 1 GB of RAM to 2GB of RAM, this report would probably take a day to complete, if at all…

The result is a 1664-page document that contains one line for person in the database, and that line is usually something like is not related to John Doe or is related to John Doe (12 steps), only for very close family members does the program list the relationship name. The list is not alphabetical and there is no index. The list is sorted by the distance in steps, but when there are a few thousand individuals at ten step, sorting these or providing an index is not exactly a luxury.

When I asked FTB to export the report to RTF, it started to recalculate the relationships. That is a broken design. As I did not fancy waiting another two hours to see what the result would look like, I cancelled the Save to RTF, but FTB did not respond to the Cancel button. After waiting more than a minute, I used the Task Manager to killed the program and restart. Task Manager reported that FTB was already using 903 MB. That is 903 MB for a 20 MB program, a 12 MB database file, and a report of, assuming a 80 bytes per individual, perhaps 8 MB text.

invalid project

When I restarted and tried to open the project again, FTB complained that the project was invalid. Luckily, it had made an automatic backup and offered to load that, which worked fine - but you have to wonder why the database was not there. I checked the database directory, and what should have been the compressed database was a file of just a few hundred bytes. That sounds like a database header without a database to me. This makes me very uneasy. Does generating a report somehow erase the database so you have to rely on a backup?

database erased

After working with the program for some time, I have to suspect that that is indeed what happened. Every time you close a project, FTB saves the entire project, even if you made no changes at all. It always loads the entire database into memory when you start working with, makes all changes in memory and then writes a new database when you close the project. So it makes sense that when the data base update logic is faulty, you do not lose a single record, but the entire database.

Problem with that theory is that I had not changed the database at all! FTB managed to lose a database while "updating" it while I had not changed anything! I am guessing that it decided to save some options and saves those options as part of the database.

I am surprised that a GEDCOM import would somehow result in an old database format.

old version

It gets weirder. FTB did not simply import the backup it had made, but stopped to inform me that the backup was saved with an old version of Family Tree Builder (huh?) or was imported from another software [sic] and needs to be upgraded. I am surprised that a GEDCOM import would somehow result in an old database format. Several different mistakes could be at the basis of this surprising dialog box, but one thing is sure: this does not engender trust in the program’s reliability.


When I click OK to save the project with the current version of Family Tree Builder, nothing seems to happen. After clicking OK and I am sure telling it to open the file more than once, it seemed to load the file, but then seemed to hang, while memory usage increased dramatically again. It let it run for a while. and after some ten minutes Windows told me that MyHeritage Family Tree Builder Genealogy Software has encountered a problem and needs to close..

When I restarted to try yet again, the file seemed to load fine, but FTB then went on a memory-consuming binge again while I waited for the program to display its menu and my data. It took several more minutes and another silly the database has been loaded dialog box that defiantly claimed the load had taken just 100 seconds, while 200 seconds was definitely closer to the truth.

kludgy internals

After this restart, one look at the screen made it clear to me how scarily defective this software is. After I had imported the GEDCOM file, the last individual in the list on the left side started with on 10... I had to resize the ID column to see 100067 and when I scrolled I saw every number from 1 up to and including 100067. After reloading it, I did not see 10.., but 16...
Surprised, I resized the ID column to see 169129. When I scrolled through the list, I found that 65.536 was no longer followed by 65.537, but by 134.599. All records seem to be in the same order as before, and a few spot check suggest that all links are still intact, but all those database identifiers have been changed for no reason at all. This once again hints at kludgy internals involving the use 16-bit integers where 32-bit integers should be used.

It is not unreasonable to suspect that this that weird renumbering is somehow related to the crash, but I when I loaded the GEDCOM into a new project, closed the project and opened it again, ID 65.536 was followed by 100.068, one more than the 100.067 individuals in the database, and numbering continued from there, to end on 134.598. Somehow, all the IDs above 65.536 magically increase by 34.532 ( = 100.067 - 65.536) between database sessions. This then, is clearly a symptom of diseased internal numbering logic.

Ancestors report

The Ancestors report defaults to showing just five generations, but you can specify another number or simple decide to show all generations. It defaults to showing photos and placeholders, but you can turn that off. You can change fonts include notes and include sources. Generation of a report takes ten to twenty seconds. My main complaint about the resulting report is that it uses consecutive numbers for the ancestors you have on file, instead of the conventional Ahnentafel numbering. I also noted that even when I choose to not display photos at all, it still shows placeholders for photos, while it should not.

more reports

The Timeline reports shows all events in the selected individual’s life. Strangely, it actually starts with the birth of the mother and father, yet omits their marriage.
The Descendants Summary is a simple line chart.


Family Tree Builder offers an Ancestors Chart, Descendant Chart, a Fan Chart and a Close Family Chart.
Like the reports, the charts are shown in the main window.

Close Family Chart

When I opt to see a Chose Family Chart, FTB produces a graphical display of 6 by 27 pages, and becomes nearly unresponsive. Scrolling through the display is a test of patience. There are various styles available, and the default produces very elongated boxes. Hoping for both a better look and a speedier, more responsive display of the diagram, I select the simple style "Basic 1". I like the resulting colours better, but the boxes are still ridiculously thin and tall. In most boxes, the lower half is empty.

When I limit the number of generations to three, the display and responsiveness improve dramatically. The boxes are still elongated but I realise this is my own fault. The first choice I made when I selected the chart from the menu was between horizontal and vertical. I just opted for horizontal without knowing what it meant. When I try again, but opt for vertical, the chart looks a lot better. I personally think that the makers of FTB swapped the meaning of vertical and horizontal, but no matter.

A three-generation Close Family Chart looks pretty nice, and I can still customise some colours and fonts, change the border style, add a shadow, change its width and colours and spice it all up with a background. A limitation worth noting is that you can only use images you already imported into FTB as a background. The way it lays out some of the lines isn’t very smart and can be a bit confusing, but otherwise the automatic layout is fine. My major complaint about this chart is that you can move the boxes, and the lines will move along, but you can not adjust the lines independently.

You can zoom out to get an overview, but even with just four generations, it is all too slow, too unresponsive to be practical. I’d recommend getting a separate charting package for larger displays.

Fan Chart

My first complaint about the fan chart is that it displays boxes, and there seems to be no way to reduce the distance between them to zero, to produce a solid fan instead of a sieve.

The Fan Chart displays pretty fast, but there is a serious limitation. Up to nine generations, everything looks fine, but FTB cannot handle anything more. It lets you you select 10 or more, but when you do so, all it shows is nine generation and four boxes of the tenth generation; two pair of parents, one for the leftmost person in the tenth generation, and one for the rightmost person of the tenth generation; i.e. on top of your paternal and maternal lines. It looks awkward.

A nine-generation overview just is not good enough, and that is true for almost everybody I know. I have eleven generations in my paternal line, many other lines are still just seven or eight generations, but some are twelve generations long, and That’s nothing special. For most Dutch children, you can not only extract a 12-, 13- or 14-generation paternal line from the records kept since circa 1650, but you can often pretty much fill up the rest of the chart from these records, and when you try do so, you are likely to encounter a few lines that are one or two generations longer. I’ve started an attempt to create a complete ancestral overview for a little girl I know, and already found one line through some nobility That’s 17 generations long now, and That’s not all that special either.

Descendants Chart

Generating a descendants chart for my oldest known ancestor in paternal line, born around 1630, takes about twenty minutes, during which FTB does not bother to show a progress dialog box, and total memory usage increases to about 440 MB. it briefly seemed to me that all the diagram has to show for all time and memory were two boxes, but then I noticed the the lines attached to these boxes, and the number of pages on the status bar. Even zoomed down to 10 %, and that is far as it goes, the application shows only part of the tree. The worst part of it all is the unresponsiveness. Although the chart is finished, the application continues to consume about 50 % of the CPU time, and still takes about half a minute to simply repaints its screen, despite the fact that it shows practically nothing. In trying to work with this chart, just zooming out to see a bit more trying to bring up the options dialog box, CPU usage regularly peaks at 100 % for stretches of ten seconds or more. The application doesn’t crash, but it cannot really handle it either. The application responds so slowly that I decide to kill and restart rather than wait for its menu to respond.

Sources and Citations

There is a Citation tab on the dialog box that lets you edit individual details, but it does not show the citations for that individual - how logical is that? It looks like it can only be used to make new citations. The Notes tab does show notes, but the citations tabs does not show citations. After some searching through the various tabs, I discovered that all the citations are present, just not in the most logical place.

To view the citations you must go the Facts tab, and click the "Fact Citations" button to bring up another, additional dialog box where you can navigate through the existing citations. You can also create news ones ,s so that kind of makes the current Citations tab redundant. I think that this dialog box should have been the Citations tab, and that it is not is only the first problem with the user interface for sourcing.

There is a drop-down box on that additional dialog box but it not there to let you select an existing citation. It merely shows the source for the current citation, yet it really is a drop-down box; it lets you select from all sources, and that makes it far too easy to accidentally change the source. When I just picked a source from the list for the current citation, FTB did not even warn me that just messed up a citation beyond recognition or repair. To actually browse through the existing citations, you need to click forward and backward buttons. More one thing went wrong when this user interface was created. All the data is there, and FTB does support citations from sources in repositories, but navigating this is amazingly awkward.

Family Tree Builder has many more features, but the fundamental issues uncovered so far make it senseless to check these out.

more features

Family Tree Builder has many more features, but the fundamental issues uncovered so far make it senseless to check these out.

One feature I just have to mention is that MyHeritage Family Tree Builder (FTB) supports import from Pearl Tree Software’s Family Tree Legends (FTL).

Family Tree Legends

Pearl Street Software abandoned Family Tree Builder before it sold the company. The last version of FTL has been a free download for some time, so if you are still using an older version, you may want to check out the upgrade.

I wanted to try a few import scenarios and see how well the FTL import works, but my experience with FTL is that it does not even install correctly. All it does it put up a messagebox that says Unable to load Family Tree Legends. Please re-install. Reinstalling or letting the installer repair does not help. It just refuses to work. It does produce a log file which reports Error: 11. If this is Windows Error 11, it is ERROR_BAD_FORMAT, which means an attempt was made to load a program with an incorrect format, and if that is the case, I can’t do anything about it, but must patiently wait for this merged company to update their download package with a working copy.

According to Cliff Shaw, the former owner of Pearl Street Software, the Pearl Street development team has joined the MyHeritage development team, which now own FTL, so updating the download with an working package should not be too difficult for them.


language and libraries

Family Tree Builder appears to have been written in C++.

Family Tree Builder uses the TE Edit Control, an RTF edit control from Sub System Inc, together with their WinPDF converter to write PDF files and apparently uses their RTF to PDF converter to create PDF files.
Family Tree Builder includes the OpenCV library (Open Source Computer Vision) for its facial recognition feature, and relies on the separate curl program to upload files.

FTB includes the OEAPI library from Nektra, but I am not sure why, as FTB does not seem to use it. It is a library to build plug-ins for Outlook Express and Windows mail, but I did not notice any plug-ins in Outlook Express.
Family Tree Builder 2.0 includes a copy of the Complete Genealogy Reporter version 2007.3 in a subdirectory. Family Tree Builder brings up the Complete Genealogy Reporter when you choose to create a Book Report. It is an unregistered copy, and therefore somewhat limited.

That’s quite a list of third-party software and components, and it not even complete. FTB includes code for zipping and unzipping its databases as well. A sizeable list of third-party stuff can be problematic, but it is not necessarily a Bad Thing. Using read-made components to build your program to avoid reinventing the wheel is a Good Thing. The major issue with this list is the absence of any database engine.



Each Family Tree Builder project uses a single database file with a *.zed extension. This is a zipped database. If you rename it to *.zip to easily extract the database itself, you do not find some binary database file, but a *.txt file. I am guessing FTB save and load operations would perform better if FTB did not burn CPU cycles to zip and unzip the file, but simple wrote it as is. It would also make it a tad easier for add-ons to read and process the file.


When you have a look at that text file, you cannot help but notice that it is a GEDCOM dialect, complete with a CHAR declaration that confirms that FTB is a code-page based program that uses Windows ANSI internally. Although FTB zips the database, FTB still belongs in the small group of genealogy programs that use GEDCOM as their native database format, and suffers all the issues associated with that design decision.

database updates

That FTB uses a zipped text file explains why FTB always rewrites the entire database. FTB cannot update a single record, because there are no records, there is only flat text. The need to always update the entire databases whenever something changes makes updates slow, and that explains why edits are not saved immediately, but only saved when you close a project. That in turn explains why FTB always saves projects on close, even if you did not make changes; it apparently does not remember whether you edited the project and is doing the safe thing.

That FTB uses a single text file as its database strongly suggests that the code is written by programmers without any theoretical background in databases, and never learned about proper file update logic either, which explains why this program manages to lose the entire database when it crashes. The portability of the GEDCOM database, is limited by the fact that is not a LINEAGE-LINKED GEDCOM as recognised by other programs, but a proprietary variation on it.


The use of a zipped text file as a database does not explain why some operations such as creation of reports, are slow. Once FTB has read the GEDCOM into its internal memory structures, it is no longer limited by the use of text file as its database, but only by its internal structures, and with all data in memory, the program should be fast, not slow.

recognising the internal GEDCOM dialect

This internal GEDCOM format is not the same as the FTB GEDCOM. This internal format contains all kinds of proprietary extensions. Many tags have "__EN" appended to them to show that the text is English, and would have a different postfix for text in other languages. These postfixes change the valid tags into invalid tags, because these proprietary tags do not start with an underscore. It is very easy to distinguish between the FTB GEDCOM dialect and the internal GEDCOM format, the internal format is not FORM LINEAGE-LINKED, but FORM ZEDCOM, which is not GEDCOM form other vendors recognise.

By the way, the presence of some proprietary tags I did not recognise in the GEDCOM header seems to confirm my suspicion that FTB saves options as part of the database, and therefore updates your databases when you change some program options.


user interface

On the surface, the user interface seems fine, but when you try to use it, you soon discover it is messed up. There are way too many dialog boxes between the user and getting a simple task done, they are needlessly complex dialog box sequences and some dialog boxes even contradict each other. Citations are not on the citations tabs, and navigating your sources is an exercise in awkwardness. To make sure the UI exceeds the scale on your annoyance meter, the program keeps throwing up rather pushy dialog boxes to obtain your data.

MyHeritage desperately needs to hire a user interface designer to fix some seriously confused dialog boxes and interface logic. And oh, it has to be one to that is willing to patiently explain to the programmers that drop-down boxes are not the hottest user-interface control available, and that adding a memory-slurping, CPU-hogging superfluous TreeView That’s so unresponsive that it is near-impossible to navigate is not a good idea.


MyHeritage’s advertising is wrong about the number of supported languages. It is disappointing that it does not support English, but only mislabels support for Amglish as "English". It is nice to see that the user-interface support is truly international, and that the entire interface becomes right-to-left for right-to-left languages.


There is no help file. None. There is a user manual in PDF format, but it is not included in the download. You need to try and access the user manual from the help to be offered a download.

GEDCOM import and log

GEDCOM import is pretty fast, but the program does not produce an import log without your prompting. When you do prompt it to write one, it turns out to be pretty useless.
FTB is one of the few program that bothers to report encoding errors, but FTB misreports its own failure to recognise a valid encoding as an error in the GEDCOM file. FTB also issues warnings for characters it cannot handle, but these warnings themselves read as if there is something wrong with characters, and do not make it clear that the issue is an FTB limitation.


Despite what its multilingual user interface would lead to you expect, Family Tree Builder does not support Unicode. FTB is a code-paged based design. Its database is a text file in Windows ANSI encoding, and its GEDCOM export supports nothing else either.

Smart Matching

MyHeritage promotes FTB by saying it features Smart Matching, but you cannot use it to compare two databases you have. FTB does not really include the Smart Matching technology, but works with the Smart Matching on the MyHeritage website.

That is not too surprising, as the website is where all the data is, but I just do not want to upload all my data including all errors, private notes and living individuals. FTB’s prompting to upload all my data, including living individuals is soliciting me to perform an illegal act, and may therefore itself be illegal. MyHeritage definitely needs to add a Private and Living filter to the upload to be on the safe side of the law.
MyHeritage should also consider offering a Smart Matching Service that I can use from within Family Tree Builder without publicly posting my data on their web site. What I’d really like though is inclusion of Smart Matching in FTB itself, so that I can compare different databases with each other..

not so smart mismatches

MyHeritage claims to perform Smart Matching whenever you edit an individual, but when challenged to perform an Easy Match that even Dumb Matching would have found, FTB’s Smart Matching failed spectacularly, by not only failing to make the match, but additionally annoying with three mismatches for other individuals.

MyHeritage claims that matching happens in real-time without slowing you down., but I experienced several minute delays during which the program was unresponsive as soon as I saved a record. I was distinctly underwhelmed by this experience.

Smart Matching may works a lot better when you do decide to upload all your data, but you will have to select a subscription plan when you do so. MyHeritage is giving away the free edition to make you pay for a subscription that unlocks its best feature. It is a good thing that the free edition allows you to check this feature out, as this first implementation of Smart Matching in FTB does not convince yet.

Reports and Charts

FTB’s reports look nice, but they do not use conventional Ahnentafel numbering, do not include indexes and take a long time to generate. When you want to save the report, FTB does not simply save, but starts regenerating the report, and you have to wait again.
The Charts look nice too, but the program cannot really handle anything but small diagrams of just a few generations. When you try larger diagrams, the program is not just slow and unresponsive, but claims massive amounts of memory and at times slows down the whole system with 100 % CPU usage.

That generation of reports takes a lot of time and consumes massive amounts of memory can be forgiven, but that your database is gone after you attempt to cancel a report save can not.


FTB defaults to putting its database in the subdirectory of the program directory, which is wrong, but the bigger problem is that Family Tree Builder is unreliable. That generation of reports takes a lot of time and consumes massive amounts of memory can be forgiven, but that your database is gone after you attempt to cancel a report save can not. That restoring that database from a backup results in less than reassuring messages about old versions, and then shows a file in which many records have suddenly changed identifiers is scarily defective. It is not defective design, but a lack of design.
The root cause of several problems is the use of a single flat text file as the project database.


MyHeritage promotes Family Tree Builder as a user-friendly, free genealogy program with multi-language support, Smart Matching and charting ability, but MyHeritage overpromises and underdelivers.

The user-interface is confused, English is missing from the language support, the Smart Matching does not work as advertised, and the program cannot really handle charts for anything but tiny genealogies. If you try to do so anyway, FTB consumes time, memory, and CPU-cycles while it becomes aggravatingly unresponsive.
FTB uses way too much memory anyway; when I load the 100k database, FTB consumes 369 MB already, That’s close to ten time the size of the GEDCOM file, while GEDCOM is an inefficient format already (so inefficient that FTB achieves a compression factor of 410 % zipping its text file, shaving 76 % of the file size).

scarily defective

FTB is a codepage-based design and will not export to any other character encoding than the Windows ANSI encoding it uses internally. It does not support export to UTF-8 or ANSEL. Its databases are not true databases, but zipped text files.

FTB does not save your edits until you close the database you are working on, so if you want to save your changes, you need to close and reopen the database. Because FTB always updates the entire database, you will lose not just your most recent change, but the entire database if the update fails. Unfortunately, FTB’s database update logic is so broken that this actually happened during my limited testing on, and it even happened on a file it was "updating" while I had not changed it at all! The program also exhibits a few seriously disconcerting indications that it uses some kludgy 16-bit logic where it should simply be using 32-bit integers.

This program is scarily defective.

GEDCOM import

time in seconds14 107s
INDI per second347,28 935,21
bytes per second75.421,07 362.611,15

product details

productFamily Tree Builder
website MyHeritage Family Tree Builder
pricefree download
requirementWindows 95 or better
noteSmart Matching is not really free
VerdictScarily Defective


download and info