Tuesday, May 19, 2009

Louisiana Slave Records, 1719-1820, The Review

An aborted earlier attempt to review this database turned into an editorial on titles. This was followed by an introduction to this database and now, today’s results of the review. Ancestry.com’s efforts to rehabilitate their presentation of this database were a waste if they didn’t make it at least as easy to use as the ibibilio.org copy of the database.

To compare the capabilities of the ibibilio.org website, Ancestry.com's old database, and the Ancestry.com new database, I had to track down some results from the old implementation.

  • Here are the search results for keyword “Falcon” from the old Ancestry.com database.
  • Here are the search results for keyword “Falcon” from the new Ancestry.com database.
  • Here are the search results for masters’ name “Falcon” from the ibibilio.org website.

To get a true picture of what the results should be, I also downloaded the original Access database from the ibibilio.org website and searched every field for “Falcon.” You can see the results here in spreadsheet format, although they don’t make much sense without downloading the key.

Here are the results of my review. To shorten references to the Ancestry.com implementations, I’ve abbreviated Ancestry.com as A.com in my review.

Feature ibibilio.org website Old A.com New A.com
Were all the 8 records with "Falcon" in them found in the search results? No. 7 of 8. Yes.
8 of 8.
No. 7 of 8.

The ibibilio.org and the New A.com databases didn’t find the case where “Falcon” appeared in the notes. The Old A.com technique of combining all the fields into one text blob, gave it the edge in this test.

Feature ibibilio.org website Old A.com New A.com
Is the original record type documented? No No No

The Document Type is one of the more important fields required to properly understand each record. I find it odd that all three implementations are missing it, especially when all 100,666 records of the original database have it. This is a big disappointment for me in the New A.com. But as it turns out, it’s not a reason to go back to the ibibilio.org website.

Feature ibibilio.org website Old A.com New A.com
All data fields from the original database have been included? No
No No

As previously mentioned, all three implementations are missing the Document type. The ibibilio.org website is also missing Name type, Birthplace (verbatim spelling), and Newly arrived (Brut). A.com’s old implementation lacks the Transcriber. New A.com lacks the Transcriber field and the Name type field, which can be an important clue regarding what part of Africa the person came from.

Feature ibibilio.org website Old A.com New A.com
All symbolic values in the original database have been replaced with meaningful text? Yes No Yes

Old A.com didn’t always use meaningful text for field names. New A.com fixes this. One nitpick is the value "housed in parish courthouses" for Document Depository. I’m assuming each parish has but one parish courthouse, in which case this should say, “Parish courthouse.”

A second nitpick would be the estate slave count, which is only included on one of the slaves in the count. This seems to be part of the original database, however. I suppose it allows the Estate slave counts to be correctly summed across all records. For all other slaves, the count can be derived from the Estate ID code (which New A.com calls Estate Number).

Feature ibibilio.org website Old A.com New A.com
The depository of the original record is given?
No No Yes
Fields are named well? Mostly Mostly Mostly

My biggest complaint with the New A.com field names concerns monetary values. Each value currently displays three fields. Sell price is described with Selling Currency, Selling Value, and Selling Value US$, like so:

  • Selling currency: Piastre = 1 P
  • Selling Value: 365
  • Selling Value US$: 365

This could be presented more clearly as

  • Sell price: 365 piastre ($365 US$)

Inventory value is worse, which appears like so:

  • Selling currency: Piastre = 1 P
  • Inventory Value: 500
  • Inventory Common Price: 500

No indication is given that “Common Price” means the same as “Value” albeit in US$. This would benefit from a treatment matching sell price:

  • Inventory Value: 500 piastre ($500 US$)

I think A.com’s choice of “Estate Number” for the Estate Identifier is a mistake. For example,
"03-F-092-047-1745" looks more like an ID than a number.

Feature ibibilio.org website Old A.com New A.com
When field names cannot be self-explanatory, their meanings are more fully explained elsewhere? No No No

These fields ought to be explained in the A.com database description:

  • Date format. Are dates mm/dd/yyyy? Why isn’t this converted to non-ambiguous format?
  • The database description ought to explain the three periods: French (1719-1769), Spanish (1770-1803), and Early American (1804-1820).
  • Do Location and Parish mean the same thing? Is Location the location of the document, the event, or the plantation?
  • Age: A.com needs to explain that if the document specified a range of years, the year listed is the mean age:  e.g., for a slave of 30 to 35 years, 32.5 was entered.  For infants, months were round off to nearest tenths of a year. BTW, it appears the estimated birth year is incorrect when a decimal point is present in the age.
  • Number of children of this slave:  It needs to be explained that this field is coded only for the mother when one is present, otherwise it is coded for the father.
  • Estate number: A.com needs to describe how to see all the slaves listed in an estate using the estate identifier. The format is LL-P-EEE-SSS-YYYY, where LL is the location code, P is the period (French, Spanish, Early American), EEE is a unique estate number for the location and period, SSS is the number of slaves, and YYYY is the year. For example, "03-F-092-047-1745" means location number 03 during the French period. The 092 is a unique identifier among the "03-F" estates. The estate contained 47 slaves and was inventoried in 1745.
  • A Birthplace of "brut"or "bozal," means the person is newly arrived from Africa.
  • Where slave came from: Specified for Africans who were born in one place and then spent time in various Caribbean countries before being brought to Louisiana.
Feature ibibilio.org website Old A.com New A.com
Fields are displayed in logical groupings? Mostly yes No Mostly yes

The ibibilio.org website places the date of sale in the document group, which I think is wrong. I also don’t like how they order of the groups. New A.com groupings look really good. However, Sold as an individual should be in the commerce group.

Feature ibibilio.org website Old A.com New A.com
Birth year is estimated from age? No No Yes

This is an extremely helpful improvement. Major kudos to A.com.

Feature ibibilio.org website Old A.com New A.com
Individuals within groups of slaves are differentiated so no two look the same? No No No

While the “Falcon” example doesn’t test this capability, A.com states that there are duplications, which would only be the case if individuals within some groups of slaves are not differentiated. This may be a nitpick. It probably doesn’t serve any genealogical purpose. But when I was first playing with the database, at times it seemed like Old A.com had lots of problems. In reality, I just didn’t understand what was going on.

Feature ibibilio.org website Old A.com New A.com
Lists of search results present the most helpful fields? Yes No Yes

Old A.com didn’t support a List or summary view at all, but always showed all fields. New A.com does have a nice List view, but on my computer in the FireFox browser, the final column is clipped.

Feature ibibilio.org website Old A.com New A.com
The search form explicitly presents important fields? Yes No Yes
The search form contains a catch-all field that searches any fields not included in the search form? No Yes Yes?

While I think the keyword field searches the fields not included in the search form, the popup help for that field re-lists the other fields. It is supposed to list the fields that are searched that aren't present in the search form.

Feature ibibilio.org website Old A.com New A.com
Search form lists all gender choices? Yes No No

New A.com needs to have Unknown in addition to Male and Female.

Feature ibibilio.org website Old A.com New A.com
Name search includes slave’s name and master’s name? No Yes Yes
Search form lists choices for Birth Location and Migration Departure Location (Origin), Lived In (Plantation Location), and Race (Racial Designation)? Yes No No

It is a real shortcoming that the search form doesn’t have drop-down lists for fields with limited values. The number of choices is relatively small, but esoteric. Without a list, it is hardly possible to utilize these field.

Possible values for Name Origin are “African,” “European,” “Could be African or European,” and “No name supplied or name is illegible.”

Possible values for Skills are beyond the scope of this article. One example is “majordomo.”

Possible values for Ship are beyond the scope of my research for this article.

Possible values for Birth Location and Migration Departure Location (Origin), Lived In (Plantation Location), and Race (Racial Designation) are listed on the ibibilio.org search page.

Passing Grade?

The new implementation of the “Louisiana Slave Records, 1719-1820” is greatly improved. Kudos to Ancestry.com for spending the time to fix-up this database.

However, in many cases the lack of drop-down lists on Ancestry.com search fields requires you to go to the ibibilio.org website to gather enough information to use the database. Many of you are going to stay there to perform your searches. If you do, you’ll be missing several advantages the New A.com database enjoys:

  • Better presentation of search results
  • Estimated birth year
  • More fields on the search form
  • Keyword search

In the end, the advantages and disadvantages of New A.com compared to ibibilio.org may be a wash.

Feature ibibilio.org website Old A.com New A.com
The grade I would give the database?   B   D   B

No comments:

Post a Comment

Note: Only a member of this blog may post a comment.