User talk:Rob Hooft

From Rodovid Engine

Hi, you wrote

I will try to play around with the localization of the nl:rodovid (I have some experience as a sysop on nl:wikipedia)

Would you like to be a sysop on Rodovid NL? --Baya 11:18, 2 March 2006 (EET)

I wouldn't mind, but I hope you do not have the hope that I will be spending many many hours. I am willing to help out to get things started a bit! Rob Hooft 20:25, 2 March 2006 (EET)

If you would like me to correct English mistakes I find on the en:/engine: sites, I may need privileges for those too....Rob Hooft 20:54, 2 March 2006 (EET)

Thank you for your help in advance :) You are sysop of engine and nl:. --Baya 23:42, 2 March 2006 (EET)

Contents 1 Do not edit engine messages into engine, only pages, help, etc... 2 New gedcom import script 2.1 Kudos 2.2 2 SOLVED 2.3 3 AUTOUPDATE/ASSIGN REMOVED 2.4 similar search OTHER PAGE 2.5 5 SOLVED 2.6 6 SOLVED 2.7 7 SOLVED 2.8 8 SOLVED 2.9 9 SOLVED 2.10 10 SOLVED 2.11 11 (age) 2.12 update & assign REMOVED 2.13 14 SOLVED 2.14 15 SOLVED 2.15 16 SOLVED 2.16 ind_refs OK 2.17 17 SOLVED 2.18 18 SOLVED 2.19 Phone numbers SOLVED 3 Father surname SOLVED 4 End of Gedcom SOLVED 5 Reimport OK 6 Question 7 uuuuuuuuuuuuuuph 8 Years with ? 9 GEDCOM in real database 10 GEDCOM localization 11 GEDCOM options 12 thanks :) 13 Rodovid Rules

[edit] Do not edit engine messages into engine, only pages, help, etc...

Edit wiki messages directly on real db, I add you to en sysop group too. Don't spend your time here. Occasionally I dump messages from real dbs and put them into language files throw all localisatins and rebuild messages into engine.

Engine - is a place for testing by users, development discussions, international sysop room, news center .... (something like metawiki) --Baya 09:56, 3 March 2006 (EET)

[edit] New gedcom import script

You try to import biggest gedcom... What do you think? I am great need comments from users... --Baya 19:15, 4 March 2006 (EET)

I have played around a bit indeed.

See: myself one of my ancestors Rob Hooft 23:44, 9 March 2006 (EET)

• Super, 11 generations!!! I know only 7. (for my children)
• More longest ancector's tree --Baya 10:06, 10 March 2006 (EET)

Here are some findings:

[edit] Kudos

Gedcom import is a winner. It makes a huge difference in the time it takes to add records. This is the most important remark, the rest is small criticism in comparison!

Thank you for your harmfull words :) and spended time.

[edit] 2 SOLVED

There are too many options. For the release on the main database, I think it is better to force people to use one-by-one record inputs.

Options: don't stop on error are not available on real db.

I was thinking about restricting it even more. But maybe if we make all the options very clear, this is not a problem.

[edit] 3 AUTOUPDATE/ASSIGN REMOVED

The option that requires all siblings to have different names is not too useful. In my family tree there are many children dying young, and the next child has the same name. Sometimes upto 4 times in a row, e.g. Person:2455

all siblings to have different names: it is not requirment, but just unresolveble problem of AUTO updating/assigning. Similar are found throw names. If one of similar have registerd father and mother current gedcom record assigned/updated to this similar. New refnum mechanism help a just in same situations.

This will indeed help if the siblings have a refnum already, so on the second input. On the first input this does not help. Also, if the birth dates/death dates are properly compared, that may help too!

• refnum help to resolve updating real db from the same gedcom file. I made this options for this case, so with ref num it is not needed.

[edit] similar search OTHER PAGE

Very important question, this mechanism can be and will be used not only during gedcom import. Let's discuss it on special page Talk:Similar search in Category:Features. --Baya 13:45, 17 March 2006 (EET)

[edit] 5 SOLVED

Maybe the Gedcom input options can be put in the "preferences" with good defaults?

Options to preferences. No. During gedcom import it is not so convienient go to self preferences to change some of options.

OK

[edit] 6 SOLVED

It should be relatively easy to import the same gedcom file with some updates again. I am using "gramps" myself to edit my family tree, and if re-importing would cost me the same amount of time every time, I might not update. Many people may be using other software to keep their data, and use rodovid as a good way to publish it.

• If the record is exactly the same, I should not even see it passing
• If the record is new, present it for verification
• If the record is changed, present it for verification

reupdateing. I add special refnum to every imported record. Script look for refnum of existing persons during importing. Unfortunatly, this checking be possible only if record imported with new version of script.

OK

[edit] 7 SOLVED

Once I have accepted a person and the import proceeds with the next, it is less useful to have the full record listed again. Especially in one-by-one import, I would like to see the NEXT record in the screen, and the previous record is no longer interesting.

I think that user must see results of his actions. But your remark is too important. Now script shown only ged_id of manually imported/assigned/skiped record instead of all context of gedcom record.

OK. I will see this next time I try it.

[edit] 8 SOLVED

If the "clan" value is copied from another field, the display should show it in the "Clan" input entry as default. Now the input stays empty.

• clan value is showing after lifetime in [] brakets.

I wasn't clear. In case of an import error, the Clan/Surname fields could be copied into the "search for person" form.

• You are strong user. :) I leave this fields blank to show additonal time that person havn’t clan & surname. But you are true I add this. (DONE)

[edit] 9 SOLVED

Is there a fundamental problem in adding new records to the search indices immediately, or will you be able to program that later?

• search indices immediately: it takes only more time. Now for every record script take from 0.3 up to 1.5 seconds ((((( it is very much. I can add option for autoindicing.

It is doing very much. But even at 1.5 seconds CPU, you can import 300000 people in one week! 6x10⁹ people takes a little longer ;-)

• Ok, )) I'll add auto indecing.

[edit] 10 SOLVED

The "I have permission" options must be changed a bit. Currently, if they are checked, the "age" field is ignored. This may be confusing.

• I have permission: maybe we must use other words to describe this option? In this case this question not for me :)

I will have a look, but I think the way it works must be changed: Currently if the box is checked, the AGE is ignored. If the box is not checked, the AGE is important. This is confusing: if an option is unchecked, what is behind should not be important. It would be better to have:

 [x] Ignore all people that were born less than [  110] years ago. '''disable this option only if
     you have permission to import information about young people in a public database'''

 [x] Ignore all people that died less than [   10] years ago. '''disable this option only if
     you have permission to import information about recent people in a public database'''

• DONE.

Much better! Is this option now set by default?

• Yes :)

[edit] 11 (age)

The age filters are used in "and" mode, and are giving "ok" if a date is unknown. Therefore all living people are imported without being filtered; only the generation of my grandparents is missing.

• age filters: How script can detect person is living or dead if script know nothing about person birthdate?

Like this:

def couldBeAlive(person):
    if person.deathdate: 
         return False
    if person.birthdate more than 110 years ago:
         return False
    return True

•  :) But in your case all young person have a "Private" birthdate value :) Unknown or absent date of death does not mean that person is living. It is not identity.

Better safe than sorry? If the person does not have a birthday nor a deathday, assume that the person could be alive and do not add it except if there is permission.

• So, you propose do not import all person for which birthday and deathday are unknown? Oldest person into your tree will be skiped.....

Indeed. Or: instead of automatically skipping the record, generate an error, and let the person judge whether it is ok to import anyway. There are ways to refine this, of course. If any other event occurs more than 110 years ago (like childbirth), this could be given as proof of being dead....

• At person import time information about person childen unknown. One way is an error. But I think this not so important. If you upload real gedcom, it will be contains birtdates, so all be ok - young people will be skiped. Gedcom with "private" dates is unreal or generated for special cause. Maybe instead of auto skipping generate an error for young peoples and for unknown dates?

Indeed, generating an error and stopping automatic import looks like the best idea. More heuristics: if a persons parents were born >160 years ago, the person is probably dead.... ;-)

• If birth date is unkonwn script generate "unknown birth date". If person is young script generate "young person (NNN)" error. If recently dead - "recently dead (NNN)".

IN RESULT: if button checked error will be generated if date is "bad" or unknown. --Baya 19:53, 15 March 2006 (EET)

[edit] update & assign REMOVED

I do not understand the difference between the autoupdate and autoassign modes.

• update and assign. Some times records into db has more info about person compare to gedcom record. In this case it will be better do not update db record but assign gedcom to existing record. So clan, surname of descendant can be taken from existing record.

So autoupdate means take all nonempty fields from the GEDCOM file and put them into the existing record, and autoassign means replace all data from the existing record by the data from the GEDCOM file?

• NO, update fully replace current record content with gedcom. Autoassign just assign id of existing record in db to to gedcom record. And it will be used for autoprocessing childs.

Important note. Script import persons from gedcom in this order.

Take a first not imported person .
while ( person not imported ) {
    if ( isset(person.father) and person.father not imported ) {
        person = person.father;
        continue;
    }
    if ( isset(person.mother) and person.mother not imported ) {
        person = person.mother;
        continue;
    }
    if ( person.father and/or person.mother and family(father,mother) not imported)
        import family(father,mother);
    break;
}
.........
person import

Autoprocessing is possible only if father and/or mother exists and both are imported or assigned.

[edit] 14 SOLVED

At the end of the import of all people and with ~80 families still to import, I saw (6x at the top of the screen) " Warning: implode() [function.implode]: Bad arguments. in /usr/home/baya/rodovid.engine/htdocs/extensions/FamilyGEDCOM.php on line 1468"

warning. I could not repeat this. Maybe this is recode already. :)

OK

[edit] 15 SOLVED

I had the young-people skip switched on at the beginning. Now I have 31 people skipped. There is no way that I can retry those 31 only, I can only remove the skipmark and start over?

• When you remove skipmark all imported record stay marked as imported. So when you start import again (without new uploading) script will go throw only nonimported records.

Great. Except that my GEDCOM cache is now empty....

• sorry (((

[edit] 16 SOLVED

Marriages come accross from my Gedcom file as "REFN" events.

• Person events list present all person events and all events of all families in which person is a parent. If family have NO events script add empty marrige event. Inyour case old script version import REFN and other refnums into events of record. So all families that your mean have at least one "event" with type REFN and empty marrige event not added. New script import all refnums into ind_refs field - so same situaton never occure.

So the marriage dates are now properly imported? What is ind_refs for? indirect references?

• I hope :).

[edit] ind_refs OK

Individual unique numbers in different system (AFN, SSN, etc....) (NB. SSN from gedcom now importing as event)

[edit] 17 SOLVED

Upon re-upload of a new version of my gedcom file, the top of the page showed a partially readable error message: Cannot open zip archive /var/tmp/phpqh7827 in ......./extensions/FamilyGEDCOM.php on line 554.

• this is unzip function warning. Already is hidden. (DONE)

[edit] 18 SOLVED

The record-per-record import is now looking much more clear than before, showing only the information to be imported. But it is very time consuming, as each record must be assigned manually, even though most records have only one candidate to update. I cannot use the autoupdate feature because of the silly duplicate names in my family.

• check "absolutely sure that....". Engine db is not a real db. Here we can do anything that we want :). Of course, all records will be doubled. But new records will be with true refnum and you can check update mechanism.

Yikes, that is too nasty to be a true suggestion.... I may do that for a small subset. ;-)

• I can delete all persons added by you to engine db.

That is a good idea! Please do!

• DONE

[edit] Phone numbers SOLVED

I get Errors: PHON - unknown structure type on a record that contains a phone number.

• ADR1, ADR2, STAE, POST added to place
• PHON - to events's ext info

[edit] Father surname SOLVED

If someone is imported with "get surname from father", but the father does not have a known surname, the error message is printed as "father surname" instead of "no surname".

• reformed.

[edit] End of Gedcom SOLVED

At the end of the import of the INDI and family records, there are still some other records left according to the gedcom status. The "continue" button is still there, but nothing is imported any more. It is probably better if the option disappears.

• Now if there are no any gedcom records to import autooptions and "continue" button are not shown.

[edit] Reimport OK

I reimported a new gedcom with a few updates and some new people, and it was correctly treated.

[edit] Question

What does rodovid mean? Rob Hooft 23:13, 16 March 2006 (EET)

• http://dict.org.ua/index.pl?word=%D1%80%D0%BE%D0%B4%D0%BE%D0%B2%D1%96%D0%B4 I hope this help you. I require too many words to describe it. :) --Baya 13:14, 17 March 2006 (EET)

[edit] uuuuuuuuuuuuuuph

I am deeply appreciative to you. --Baya 16:39, 7 March 2006 (EET)

Don't. I am a programmer myself, I know how important it is to get feedback from first time users. My comments will not be as useful the second time around.... Rob Hooft 22:17, 7 March 2006 (EET)

Should I try to re-import my GEDCOM? Rob Hooft 22:47, 7 March 2006 (EET)

Yes. Refnum is very usefull and never disapear. I will answer for all on thursday --Baya 01:42, 8 March 2006 (EET)

[edit] Years with ?

One of the birthdays in my Gedcom was 179?-10-27. This was interpreted as a year of 179, and extra info ?-10-27; It was meant to be that the units of the year were unreadable in the original document. Rob Hooft 16:22, 18 March 2006 (EET)

• This is bug. 179?-10-27 - is a wrong gedcom data format and must be imported into extrainfo. Treu format is 27 OCT 1790?. Recoded. 179?-10-27 will be added to extra. 27 OCT 1790? will be converted to 'ABT 27 OCT 1790'. --Baya 14:11, 20 March 2006 (EET)

[edit] GEDCOM in real database

You ask whether it should be enabled. I would love to import my family tree into the real data base. Rob Hooft 23:43, 20 March 2006 (EET)

[edit] GEDCOM localization

The GEDCOM error "unknown birth date" does not come from a MediaWiki message I can find in Rodovid NL

• status and error messages are hardly coded. I move then into messages over some time. Now I divide parser code between absract and real classes (I want add couple parsers for other genealogical data forms) --Baya 23:47, 21 March 2006 (EET)

[edit] GEDCOM options

The option "preview every record before import" should probably default to on. I started on NL importing my tree, and I noticed the surname from was set to "this is an error". I left it like that but when I came back it was set to "taken from father".

• old bug. recoded. ignore and check every are checked on new upload. --Baya 23:44, 21 March 2006 (EET)

[edit] thanks :)

I'm so ill from yesterday ((( Waked up from bed to read last news...

[edit] Rodovid Rules

(Message sent to all Rodovid sysops)

Hello

There was a pretty large discussion on Rodovid Engine about establishing rules. You can see the result of this discussion on this page:

• Rodovid Rules

We propose to begin applying these rules Sunday, August 1 and to watch how it runs. Sincerly. --Christophe Tesson - ^talk. 12:10, 24 July 2010 (EEST)