GeneWeb - Frequently Asked Questions

At the first use of GeneWeb, I get the message: Cannot access file "start.txt".

Look at the doc: "Two frequently encountered problems". If it does not work better, please tell me.

It does not work under Windows!

The first thing you have to check: is the gwsetup program which you launched still up and running? Because, for it to work, it is necessary that gwsetup remains. If this program were running in a window and if you closed that window, it cannot work. If it is your problem, double-click again on gwsetup.

Later, for the program gwd which allows you to browse your database, same issue: you launch it by double-clicking on it and you must not stop it while you want to navigate in your database.

It may happen that Windows does not recognize "localhost". In this case, try the address "" instead:

Otherwise, a frequent problem is that your Web browser, seeing that you entered a Web address (http://...), tries to connect to the network and if you disconnected your modem, it refuses. It is an error of Windows, because the address being "localhost", it should not try to connect: "localhost" means "your own computer".

It is due to the fact that your internet provider has installed a program for automatic connection on your computer. This automatic connection is responsible to this phenomenon. A solution is therefore to remove it. Search in a menu, which must be named "preferences", "options" or "configuration". There you see the automatic connection and can remove it.

A different solution is to use another Web browser: if your internet provider installed Internet Explorer in your computer, use Netscape. If Nescape were installed, use Internet Explorer.

A third solution is to connect your modem, and wait for the connection. When it is done, does it work? If yes, then we are sure that it was the problem. Then disconnecting your modem, it often will continue to work. Of course, it is not very satisfactory since you have to connect each time you want to use GeneWeb, even if it is for only 10 seconds.

Information from an user (Paul Laredo, thanks to him):

  1. launch gwsetup (or gwd)
  2. launch your Web browser
  3. stop the automatic connection with the provider
  4. type the address of the service
  5. it displays that the page is not available off line, you must not click on the button "work off line" which is displayed but on "connect".

My genealogic software tells me that the base holds 3,000 individuals, but GeneWeb says: 3,050 individuals. Why this difference?

In GeneWeb, parents are always recorded in couple. When there is just the father or just the mother in the GEDCOM file, GeneWeb adds the other parent, giving her or him a question mark as first name and surname. The difference are these missing (but having existed) parents.

When displaying relationship links, some are missing.

To avoid that the listing of relationship links be too long, GeneWeb deletes some of them when displaying the result. For this, it uses the following method: when a relationship link is found with a common ancestor, all the ancestors of this ancestor are eliminated from the research.

A consequence of this pruning is that if one asks for the computation of a relationship link with an ancestor, he will be displayed only as an ancestor, even if he can sometimes also be a cousin.

For example, if my father and my mother are cousins, and if I am looking for my relationships with my mother, it only indicates that I am her son, not that I am her cousin's son, even if this link really exists.

Pruning is necessary: without it, for families a little bit "intermixed", the displaying rapidly becomes invaded by a lot a relationship links and eventually not understandable (I tested it).

The non displayed relationship links are however real relationship links, and are counted in the consanguinity computing.

In CGI mode I just get a message "Internal Server Error"

A frequent problem is about access rights. The Web server often runs CGIs as an user with very few rights (typically user "nobody" under Unix). It may not be able to run "gwd". Check the access rights of "gwd" and all the GeneWeb files!

Why is the age of persons of more than 80 years old not displayed? Why aren't they in the statistics of the oldest persons?

In GeneWeb, the indication of death has three states: "alive", "died" and "don't know". The age is displayed only for "alive" (still living) persons. If the age is not there, the reason is that either the person is dead (without indication of date), or the indication is at "don't know".

If you added this person with GeneWeb and you did not specify if he/she is alive or dead, or if your database comes from a GEDCOM file (the indication "still living person" does not exist in standard GEDCOM 5.5), GeneWeb applies the following algorithm:

You have to fix these indications if they are incorrect.

The persons having ? (question mark) as first name and/or surname seem to have a specific status. Which one and why?

These persons are considered by GeneWeb as "dead-ends": if they are children, they cannot be spouses, if they are spouses, they cannot have parents. And they cannot have associated notes.

Moreover, if they are spouses, they are not exported in the GEDCOM files and their personal information (if any), is consequently lost.

It is sometimes possible, by hacking, to be able to give them notes and to make links, but a warning: at the first cleaning up of the data base, these notes and links are going to disappear: it is therefore advised not to do that.

The first name/surname "?" being therefore "reserved" in GeneWeb. You can't name persons to "?", if you want them to be linkable to the rest of the database. Also if you want to attach notes or relations, (godparents, adoptive parents or children, etc.), to the persons, you can't name them "?". The solution is to edit these persons and change their surname or their first name to something else: put "N", "Ne", "N...", "x", "??" or what you want but not just "?".

Could you improve the "advanced request" by "this" or "that"?

The "advanced request" is a hack I programmed but which is not very satisfactory. As GeneWeb does not use a real database system, any improvement of this part would consist on rewriting a new database system, which I don't plan to do, because it is a huge work.

If, one day, GeneWeb uses a true database system (it is among the possible directions), this advanced request will be probably rewritten, and will be much more powerful and easy to use.

Would it be possible to have a complete displaying of ancestors, even with repetitions of Sosa numbers?

I understand that it would be useful, but there are big problems. I already experimented with this feature but I became aware, for persons having many ancestors, of the fact that the displaying was full of references to other ancestors and became unreadable. Moreover, this triggers much computation ("exponential algorithm").

If you want to know if an ancestor line is full, look at the reverse problem using "missing ancestors".

What is the convention of short dates displaying?

     Smith 1935-1950 = born in 1935, died in 1950
     Smith 1935-     = born in 1935, not dead (still alive)
     Smith 1935 +    = born in 1935, died unknown date
     Smith 1935      = born in 1935, don't know if dead or not
     Smith +1950     = died in 1950, unknown birth date

        /1935        = before 1935
        1935/        = after 1935
        ca 1935      = about 1935
                       perhaps in 1935
        1935/1938    = in 1935 or in 1938
                       between 1935 and 1938
This convention does not refer to a specific "standard". If you know a better, more known, more standard convention, don't hesitate to tell me. Thanks.

Would it be possible that you add more information about persons in the "families" update forms, in order to go faster?

It would certainly be nice, but I don't like adding things. The more I add, the more people ask me. In the beginning, I put only the person's name, then I was asked to add dates, I added dates, then places, I added places, then people ask me to add the dates precisions, the occupations, and so on. All these informations increase the size of this "family" form which is already rather complicated.

I know that it is tedious to have to return to the "person" form for each added person, but I am not sure that adding and adding things be the good solution. It is also an intrinsic problem of HTML, which is not very "dynamic". However, I am personaly against solutions using Java (risks of security holes) and JavaScript (language actually destinated to be invaded by advertising, among other drawbacks). Other solutions are being studied, but not easy to implement.

In an update form, I filled in some fields, but they disapeared

Some fields are taken into account only in certain circumstances. So, if a person is denoted "alive", the fields "date" and "place" of death are ignored. If, in a "family" form, the field "separated" is selected, the field "divorce date" is ignored. If the indication of a person is "Link" (instead of "Create"), the fields about "sex", "date" and "place" are ignored. And so on.

Generally, in such kind of forms, the choices are invalidated or validated back when one select such or such button. It is not possible to program that due to the fact that GeneWeb is limitated to pure HTML: JavaScript should be used, which is against my principles (see the question about the families update forms).

When I make a search by place/surname, if the result give several persons for a surname, only the first one is accessible

Indeed. In a first version, I made a link to the surname itself, and not to a person, but when the surname holds many persons, you don't know at all who is concerned by this "place" and it is difficult to find them.

To give the first person is more "informative". Yes, it is incomplete, and the ideal would be to have the list of all the persons concerned by this place, but it supposed that I programmed several things more, and it has the drawback that the request uses much more memory space.

You can access the list of all the persons of this surname by clicking on the surname in the person's displayed page.

When I browse my database, I get sometimes the message "probable loop in the database; somebody is his own ancestor"

Your database probably comes from a GEDCOM file (it cannot happen with a database created under GeneWeb). At import time, this error has been detected but the database has been build. However, some functionalities cannot work, like the relationship computing or the ancestors display.

To fix this problem, you must know which person or persons are involved. Create your database again from the GEDCOM file (if you made changes in your database, you can create it with another name to avoid overwriting it). If you use interactive commands, look at the result of "ged2gwb". If you use gwsetup, after the import, go to the main menu and click on "traces of the latest command". The persons who are their own ancestors are displayed with the message (in English): "So-and-so is his/her own ancestor".

I typed "Lemieux" and it answered me "Louis Maheux"

GeneWeb uses a spelling correction which can sometimes give strange results, indeed. This correction applies to first names, surnames and complete names.

When you type a name, GeneWeb searches if it exists like you wrote it. Otherwise it looks for approximated answers. For that, it simplifies the name with the following rules:

So, the first name "John Charles" has the signature: "JN CRL", which is the same as "Juan Carlos", "Jean Charles" and "Johann Karl". If you search for "John Charles of Bourbon", you can find "Juan Carlos de Borbón" (particles "de", "of", "von" being suppressed also).

For the example given as title, "Lemieux", the signature is "LMX" which is the same, except for the space, as "Louis Maheux": "L MX". If you want a research by surname, click on the button "surname" before clicking on "ok".

Why the fields "source" in gedcom are not taken into account?

Actually, the sources are taken into account only if they are texts, because GeneWeb has not planed to record them differently. If they are represented by a structure, GeneWeb does not know how to treat them because it does not know what are the meanings of the structure items.

In forms, the "tab" key seems to jump from a field to another in an unlogical order

It is a browser issue. There is no way to ask them to have a correct behaviour. Some browsers work correctly, others no.

Return to Directions for Use

© Copyright 2001 INRIA - GeneWeb