Modern Software Experience

2010-12-06

genealogy application

StamboomNederland

I wanted to like StamboomNederland. Many years ago, I asked for it.
About eight or ten years ago, I suggested to the Central Bureau of Genealogy that they should support digital genealogy like they support paper genealogy. Keep copies of it and let patrons use these genealogies. Since then, I have repeated that suggestion more than once. So I was thrilled to hear that they had started a project to do just that. I took the time to attend two presentations about it. I treated the release as a major event, being quick to release first impressions, but taking time to review each aspect of the product.

I still want to like StamboomNederland, but I can't. I asked for it, but it did not ask for this. As I examined one aspect of StamboomNederland after another, I was repeatedly disappointed by StamboomNederland's failure to do even the most basic things right. It became increasingly clear to me that StamboomNederland is hardly functional.
I initially tried to explain some issues away by noting that it had been rushed out to meet a deadline. However, in the past few months, StamboomNederland seems to have improved little, if at all.

user interface 

The graphical style isn't very consistent and if it did not have the CBG logo, you wouldn't recognise it as a CBG project. Lots of things you want to do are not on the menu. Some things can only be done by clicking the unintuitive icons. Other things can only be done by navigating to a button on some less than obvious page. Random buttons are in ALL-UPPERCASE. The StamboomNederland user interface is a randomly slapped together mess.

privacy and copyright

A major issue with StamboomNederland is that it does not respect privacy. You do not get a username, StamboomNederland uses your email address instead. When you make a project public, StamboomNederland makes your email address public.
You can delete projects, but you can not delete GEDCOM files you've uploaded. Even mailed request to delete these files are not acted upon. Even a direct mail to the project manager did still not result in deletion of uploaded GEDCOM files.

Users have complained about a copyright issue. The StamboomNederland Terms & Conditions state that once you haven't logged in for five years, ownership of your database automatically transfers to the CBG. Despite complaints about this data-grab, the terms & conditions have not changed.

Having to use StamboomNederland should be classified as cruel and unusual punishment.

slow

Saving your changes is not near-instantaneous as you would expect. Saving even a tiny bit of information can take multiple seconds, even minutes. Saving your changes is unacceptably slow.

using StamboomNederland

Trying to use StamboomNederland is exercise in patience and determination. Having to use StamboomNederland should be classified as cruel and unusual punishment.

StamboomNederland is raw and unpolished, full of unexpected limitations. Fully-specified place names are limited to a fixed size of just 50 characters, thus forcing users into the bad practice of using abbreviations instead of full names. Dialog boxes do not fit the browser window. You can enter a marriage date, but not a marriage place. Sometimes the graphical display works, sometimes it does not. When it works, it does not show patronyms. Although it is perfectly normal to record stillborn without a first name, StamboomNederland demands a first name for every individual you enter. This kind of inflexibility already makes it unsuitable for real-world genealogy already, but things are worse.

consistency checks

StamboomNederland does not have any consistency checks. None whatsoever. It does not warn you about impossible age differences between parents and children. It does not warn you about death dates that are before the corresponding birth date. It does not even warn you when you add four mothers and five fathers for the same individual. StamboomNederland is a dumb relationship editor that allows every possible and impossible relationship. A relationship editor isn't a genealogy application.

GEDCOM support

The buttons for GEDCOM import and export are not where you expect them, but that is the least of the GEDCOM problems. One issue that caught several users of guard is that StamboomNederland's GEDCOM import does not replace the data in the existing project, but adds to it. Several users found themselves with lots of duplicated data and now easy way to remove it, because the default project cannot be deleted.

StamboomNederland does not feature native GEDCOM import or export. It relies on the raw XML input and output of the database system. StamboomNederland's GEDCOM support is built on top of that; GEDCOM import is done by converting from GEDCOM to raw XML and then loading that, GEDCOM export is done by dumping the database to raw XML and then converting that to GEDCOM. That approach is good enough for a proof a concept, but not for a production system.

memory inefficient

The GEDCOM import not only happens in two steps instead of one, the GEDCOM to XML conversion is also remarkably slow and memory hungry. The CBG did not publish the specifications of their new server, but it seems reasonable to assume that it has more than 4 GB of RAM. While desktop applications manage to import it on machines with just 1 or 2 GB of RAM, StamboomNederland choked on HundredThousand.ged. If the server has just 4 GB RAM, and that is a rather small amount for a brand new server, then a back of the envelope calculation shows that Kensas's GEDCOM to XML conversion needs more than hundred times the size of the GEDCOM file it converts.

However much RAM the server actually has, the GEDCOM to XML conversion process is so greedy and inefficient that there was no memory left for processes. After this experience, the CBG decided to place a limit on the GEDCOM upload to avoid it happening again. Kensas hasn't fixed the GEDCOM to XML conversion yet, so that limit is still in place.

slow

The GEDCOM import does not produce import log file. It does produce an import process action report. This is largely useless, but it does contain time information. This allowed to establish that the import of the 1 MB GEDCOM file took 71 seconds, which equates to a pathetic import speed of just 68,48 individuals per second - on brand new server hardware!
Most desktop genealogy applications achieved better performance on my old Windows XP machine. For example, PAF did it in 11 seconds and RootsMagic 4 did it in 6 seconds. That PC is about seven years old, and StamboomNederland takes almost twelve times as long on brand new server hardware.

StamboomNederland demands payment the moment you decide to take your data elsewhere!

GEDCOM export

StamboomNederland's GEDCOM export is a paid feature. In other words: StamboomNederland demands payment the moment you decide to take your data elsewhere! Even if StamboomNederland were a fantastic product, that fact alone would be enough for me to disrecommend it.

GEDCOM quality

The quality of the exported GEDCOM leaves a lot to be desired. There are a few small things that StamboomNederland gets rights. Some of the complaints about StamboomNederland are caused by defects in GEDCOM produced by other applications.
Still, there are two major problems with the quality; StamboomNederland has an utterly confused notion of sources, and as a result destroys all your carefully recorded sources to bluntly replace them with the name of the GEDCOM file (as if the GEDCOM file that contains your sources is the single source), and add notes for many bits of GEDCOM it can't handle. By the way, those notes aren't simply unhandled GEDCOM tags, like several applications do, but the raw XML equivalent of those GEDCOM tags. StamboomNederland's GEDCOM support is seriously messed up.

search

StamboomNederland's search is a prototypical example of a checkbox feature; it has been hurriedly slapped on, so they can tick the search box on the feature list. It has a search feature, but that is all it has. There are no search options. None at all. You cannot even specify that you are searching for a last name. The simplistic search will always search all fields. And oh, that includes all those notes fields containing raw XML…

paid features

The paid features are disappointing. GEDCOM export should be free feature. The so-called collaboration feature is merely a multi-user feature and the management of project users does not seem to work. The media support is limited to PDF and JPG files. The report feature does not show reports online, it merely creates a PDF or RTF file that you have to download and view on your desktop.

This is not a product with a few defect that needs to be fixed. This is a bundle of defects presented as a product.

untested

Previous articles in this series discuss these and other issues in some detail. Some issues are more serious than others, and a few may have been missed in testing, but it impossible to miss all of these issues. It is safe to say that if any serious testing was done, most of these issues would have been found. StamboomNederland is not reliable at all, the Interne fout (Internal error) page is the one you'll see most often.

showstoppers

Quite of few of the noted issues are showstoppers; issues that are so serious that they are reason enough to postpone introduction until these issues are solved. That this mess was released as a product is reprehensible. The overall quality of StamboomNederland is below that of a quick & dirty prototype.
This is not a product with a few defect that needs to be fixed. This is a bundle of defects presented as a product.

competition

promotion

StamboomNederland is a CBG product sponsored with tax money, and the CBG is already known as the national repository of genealogy information. The CBG published a four-page article about in their quarterly magazine. The same issue contained another four-page article about Verborgen Verleden (Hidden Past), the Dutch edition of Why Do You Think You Are. The verborgenverleden.nl domain leads to the StamboomNederland site, and this domain was advertised during all the eight episodes of Verborgen Verleden.

StamboomNederland is so bad that, if it did not have a CBG logo, the CBG would warn against using it.

unfair competition

The CBG got the money to make Stamboom for free, has the advantage of being of a semi-official government organisation, and got free advertising too. No wonder then that existing Dutch genealogy sites considered StamboomNederland to be unfair competition. I say considered, past tense, because StamboomNederland is a completely failure, and not one existing genealogy site is worried about it anymore. StamboomNederland is so bad that, if it did not have a CBG logo, the CBG would warn against using it.

complete failure

The StamboomNederland site lacks a statistics page that shows you many users and profiles it has, but I feel pretty sure that if it did, you would see an huge initial interest because of the cited reasons, quickly dwindling to no more than a trickle of incidental visitors.
Anyone who tries the site is put off by its ugly design and awkward, illogical menu structure. Those who persevere regardless are confronted with breach of privacy, its completely unworkable, utterly confused ideas of sourcing, tjhe slow and defective GEDCOM support, diagrams that do not work, and its inability to work with large genealogies. StamboomNederland manages to get practically everything wrong. StamboomNederland is a completely failure.

It is high time the CBG acknowledges the deplorable truth; StamboomNederland is a complete failure. Every day that this monstrosity remains online hurts their reputation. It is time to face that fact and pull the plug.

not salvageable

The site that Kensas created is so poor and so messed up in so many ways, that it does not seem salvageable at all. Moreover, the lack of any significant improvement since its introduction suggests that even its creators are already unable to find their way through the slapped-together mess they created.

replacement

Kensas clearly underestimated the project c.q. overestimated their own capabilities. During the official presentation they were crowing that their StamboomNederland is fast and efficient, and could handle anything. Verily, the CBG would even be able to add additional event types without any impact on performance.

The reality is completely different from Kensas's grandiose claims and promises. StamboomNederland runs on a brand new server that the CBG bought for the purpose, yet soon after introduction the CBG had to restrict the GEDCOM upload to medium size genealogies, because StamboomNederland is incapable of handling larger ones.

It is high time the CBG acknowledges the deplorable truth; StamboomNederland is a complete failure. Every day that this monstrosity remains online hurts their reputation. It is time to face that fact and pull the plug. The CBG should get their money back and contract a more competent company to develop a complete replacement.

ready-made products

The CBG could hire another company, but one has to wonder why the CBG wants to develop a web application for genealogy at all. Several individuals and organisations have already spent years developing genealogy web applications and provide these ready-made products for free or some reasonable license fee.
It is not clear why the CBG wanted to created their own product when they could have used or adapted an existing one. The major web genealogy products certainly allow customisation of the user interface to fit the house style. The open source products even allow modifying the functionality to suit any specific demands.
Perhaps they wanted something unique, that is all their own. The best way to accomplish might have been to buy Zooof before MyHeritage did. StamboomNederland is unique in ways it should not be.

e-depot

editable

A more fundamental issue is that it is debatable whether StamboomNederland should be an online genealogy application at all. There is no doubt that products like TNG and phpGedView are more flexible than static pages generated by some desktop application. It sure seems desirable that on-line genealogy be editable, but that still does not mean that it is the right solution for everyone and everything.

StamboomNederland was not conceived as a genealogy application. StamboomNederland was conceived as a digital depot, a place to safe-keep genealogy research for future generations. Recognising the various Dutch GEDCOM dialects correctly is way more important than offering editing capabilities.

digital depot

One of the things the Central Bureau of Genealogy does is manage a depot of  genealogical publications. It collects, stores and indexes the publications, to make these readily available to its patrons. Many researchers offer a copy of their genealogical publication to the CBG. It has become common for researchers to offer a disk containing their genealogical database along with their printed work, but the CBG did not have a depot for digital genealogies. In the past, they experimented with Pro-Gen databases linked from their catalogue, but it was not ideal.

electronic depot

Patrons asked the CBG to provide a depot for electronic publications. If the CBG were to offer an electronic depot, and if that e-depot would understand and display GEDCOM files, then patrons could upload their data themselves. Moreover, users can update it themselves, so that they do not have to bother the staff for every corrrection or addition.

An e-depot for genealogy would not only be a great place to store and present published genealogies. It would also be an ideal place for researchers to store current research while they are not ready to publish yet, just to make sure that their research is preserved. All that requires is the ability to keep uploaded data private.

server

The CBG presents StamboomNederland as its e-depot. The StamboomNederland server is housed  in the server room of the National Archive and is included in the National Archive's backup regimen. It almost seems an insult of the National Archive, but the CBG makes its own backups as well. All this to ensure that genealogies stored in Stamboom Nederland are safe from digital disasters. Sounds good, right?

Placing the StamboomNederland server with the National Archives doesn't make it an e-depot.

It sounds good, but it isn't. StamboomNederland isn't an e-depot. StamboomNederland is an extremely poor genealogy webapp, and even if it were a great genealogy webapp, it still wouldn't be an e-depot. An online genealogy application isn't an e-depot. Placing the StamboomNederland server with the National Archives doesn't make it an e-depot.

e-depot features

StamboomNederland is a poor genealogy webapp that makes a mess of your GEDCOM file, your sources in particular.
A real e-depot would not mess anything up, but simply respect all the information in your GEDCOM file. A good e-depot would offer versioning. A great genealogy e-depot would provide you with a report on the inconsistencies and implausibilities in your database. A smart e-depot would offer a smartmatching function, matching your data against all public databases. A truly great e-depot would offer a programmatic interface so that users of supporting genealogy applications can safeguard their data from within the application they use.

The CBG wanted a single system that would be several things at once, a system that would be everything to everyone. StamboomNederland is nothing to no one.

forgotten

The CBG started with the idea to build an e-depot; an electronic equivalent of the paper depot they already have.
What seems to have happened is that they lost sight of their original goal. They somehow got enamoured by the idea that it should be possible to edit the content of the e-depot, and that idea then morphed further into no more than an ordinary genealogy application running on a server placed with the National Archive.
The CBG wanted a single system that would be several things at once, a system that would be everything to everyone. StamboomNederland is nothing to no one.

Creating a simple online genealogy application for beginners may be a good idea, but an e-depot and a beginner's application are two different things. Trying to build a single system that is both isn't a particularly good idea either, as beginner's research isn't likely to be worth preserving.

The e-depot idea seems to have been forgotten. In its own article, the CBG writes that it has plans to extend the genealogy application's functionality. What the CBG should do it is provide the e-depot functionality that Dutch genealogists are still waiting for.

links