VmWareI have finished entering all of my lamb growth data for both 2009 and 2010 into the National Sheep Improvement Program’s (NSIP) new database tool, Pedigree Wizard. NSIP moved this year from being supported by university researchers to a commercial company in NZ Australia called Sheep Genetics, makers of LAMBPLAN. [Some info from them on their origin: Sheep Genetics is a joint program between two Australian industry research companies that represent the red meat industry (MLA) and wool industry (AWI). As such, most of the development for LAMBPLAN and Pedigree Wizard comes from Australian grower levy funds. While Sheep Genetics is operating under cost recovery for basic services, it is not commercial.] This is my first year enrolling in NSIP and I am looking forward to seeing what the metrics tell me. Here are my impressions of the new software. With apologies, because I’m a software engineer,  and I live, well, here, in the birthplace of Microsoft and Amazon. So I review it from the spoiled perspective of someone who usually gets to use the latest and greatest software.

LambPlanScreenFirst off, I haven’t used software that looks like this since I was twelve. Yes, it looks like it was written in the ‘80s, in the era of DOS. Before the widespread adoption of the computer mouse. Do you remember such a time? My first clue that this may indeed be the case was that there are some panels, such as the above, with no click-able buttons on them. This startled me for a second, there is something disconcerting about being in a screen with no buttons, like Alice finding herself in a room with no usable doors. But then I resurrected my pre-teen memory of what to do. To get out, you have to hit <enter> as many times as it takes to traverse all the fields, and then the panel will close. Or <escape> sometimes works. The color scheme also hints at the days before UI and UX regularly fell off the lips of HR recruiters. And the fonts! Soooo retro!  Winking smileThe .dbf files it generates hint that it’s written in some form of dBase, which I’ve never used. I think I’m too young, just a GenX babe who only ever learned SQL.

No Go on the 64-Bit

The database will not run on 64-bit machines. Will. Not. Run. I disregarded this advice at first, thinking surely I could overcome with the Windows 7 tricks of run-as-XP and whatnot. No can do, even Windows 7 tells you, post-install, this will not run. I had to download VMWare’s free virtual machine VmWare Player, load an old copy of XP into it (thank goodness I hadn’t cleaned my shelves lately of outdated software!), and install the app from in there. This works ok, it was straightforward to set up and to save files out to the shell OS. But this may feel too intimidating for many users, and quite a few on the discussion list have piped up saying they have 64-bit machines and had to find alternatives. The official advice from NSIP if you have a 64-bit machine: buy an old computer just to run Pedigree Wizard. Yuk.

I’ll Just Consult the Manual

There is a help manual, last updated in 2002, but it’s not very helpful or thorough or modern. And it has no specific advice for our breed’s conventions, it’s very generic. I really had to scrounge to find answers to some of my questions- through NSIP discussion groups and asking questions of our NSIP leads. The most hilarious piece of advice in the manual is this:

A backup to a floppy disk should be stored away from your computer, maybe in another location or building.  When storing floppy disks wrap in tin foil and place in a plastic bag for greater protection.

Forklift Upgrade On Horizon?

Sheep Genetics promises sometime soon, they are re-writing this baby from scratch. Well, for a thirty-something software application still living at home, hey, no rush. All kidding aside, it’s not like the sheep industry has millions of potential users all clamoring for a big software company to write them a fancy app. Beggars can’t be choosers, I suppose. In some ways, it reminds me of something a grad student would have made, just a no-frills, functional piece of software; robust, but not pretty or convenient for the uninitiated. And maybe that’s its origin.

EntryWIndowCan You Parse 16 Character Strings?

Entering the data is a little tedious. You can see the style of sheep IDs above: 64 is our breed code, 0075 is my farm code, then you add the 4-digit birth year, and have 6 digits left for the sheep’s identity. On some screens, they break it up for you, on others, it’s all munched together and your eyes have to pick out the components. It’s a little dizzying if you are looking at a whole list of them. The KHSI convention has been to use the registration ear tag string, dropping any extra numbers beyond three digits. This could be a problem for people who have more than 999 lambs in a year, but that’s apparently rare or nonexistent in our registry. I have one ewe with a longer number, JPS60224, so it seems a little weird to truncate that, but I had to. The KHSI community agreed they wanted to keep using three of the six characters for the flock code, so that people can read the numbers and know whose sheep they’re looking at.

What to Put In

I originally had visions of some massive data entry task to get started for the first year. But it turns out, you only have to enter the sire and dam of lambs, not the entire pedigree. If their sire or dam was in the NSIP before, you have to investigate to find the animal’s original NSIP ID. If it’s a non-NSIP sire or dam, you enter them with your own flock code, and a 0000 for the year inside the ID string. This apparently “tells” Pedigree Wizard that this is a reference animal, because it doesn’t include these animals in the browse windows where you can add weight, pedigree and other data for them. The other alternative is to not specify them at all, as apparently the most important data is the growth data, not the ancestry linkages which have no data.

I found it was easiest to enter all my lambs and their basic data first- birthdate, sire/dam, birth weight, birth and rearing type, sex and lambing ease; and then go back and enter their 60 and 120 day weights and other info about them in subsequent passes. It is really easy to make typos in this interface and hard to spot them. Fortunately it has a decent error checker and validation tool you can run to help you find the mistakes. I realized right at the last minute that it’s worth asking the error checker to search all records, not just those for my flock ID or the year in question. I had a couple of sheep where I’d typed the wrong breed code or birth year, and I didn’t find them until I let the error checker check everything.


To Export: Click Heels and Repeat There’s No Place Like Home

Once I was satisfied I’d entered all my data correctly, I was a little stumped that the Export Data button was grayed out, and nothing I tried would enable it. Back to the manual. It says:

Do not export data using the EXPORT DATA button on the EXPORT AND DATA VALIDATION screen.  This is because there is an easier way of exporting data.

An easier method than pushing a button? I’m not sure what could be easier- maybe just wishing the data would export and it would obey my mental command? I don’t know, because the manual never says. I finally figured out that there is a harder way to export the data, and I think it’s the only way: you just make a backup zip file, and email it to database@sheepgenetics.org.au. They apparently email you an update file back that contains your EBV (Estimated Breeding Value) data, and you import that back into Pedigree Wizard to browse your numbers. Ok, so I’m a little disappointed at the tempting but grayed out button for exporting, but I guess email is still alright.

Database Went on Holiday

Data submissions are processed after the 1st and 15th of each month, so the turnaround for information is supposed to be fast. I hustled to finish my data entry by the end of December, only to be disappointed today to see an email announcement today that the guy who does the data processing is on vacation. So he won’t process our submissions until the end of January. <sigh> Oh well, I can wait.

My data probably won’t be very revealing this first year anyway. I have a few sheep that were already in NSIP or trace back to it, so I will benefit a little from that extra data beyond my own. But the rest of my sheep are just going to reflect their own growth curves based on dam age and birth type, and I’ve already done a lot of analysis on that. In theory, it won’t be until I’ve been in NSIP for a few years before it’ll start giving me cumulative data that would be impossible to hand calculate.