Attention all hands

[Randi]

Are there any macOS users currently working on the project?
Please contact me.

[Michael]

I've been transcribing Haida 1943 and I thought it would be useful to add Lying to as a sub-type to Sailing Info in the Event Wizard. I also noticed that the sub-types Location and Lying to both start with L. Also, the Event types Sailing Info and Ships both start with S. If the types and sub-types start with a unique letter, then just entering the first letter will bring up the unique value. For the case of Sailing Info and Ships, just entering S brings up Sailing Info, but Sh is required to select Ships.

Similarly, for the sub-types, L will select Location but Ly is needed to select the new sub-type Lying to.

My question is this:

1. Should I leave things as they are, so you need to enter Sh for the type Ships and Ly for the new Sailing Info sub-type Lying to; or,
2. Should I change the type Ships to Vessels and the Sailing Info sub-type Location to Places so one letter will suffice for all types and sub-types?

Note: The type Other has Aurorae and Astronomical as sub-types, but they are used so rarely that I have no intention of changing them.

Do you prefer to have no change to the Event Wizard with the occasional need to use two letters to choose your type or sub-type, or do you prefer a small change to the Event Wizard and then need only one letter to get your type or sub-type?

[Hanibal94]

I would prefer no change to the Event Wizard with the occasional need to use two letters to choose your type.

But what's the difference between "Hove To", "Stopped" and "Lying to"? I think I may have been using "Stopped" for all of them - I was definitely using it for "Lying to"...

[Michael]

They're all the same and using just Hove to or Stopped for those cases is perfectly fine. Similarly, Moored and Anchored are the effectively the same as each other, but they are different from Hove to, Lying to and Stopped. For the first case, the ship is attached to a dock or the ocean floor, so it's not moving. If the ship is stopped, hove to, or lying to, it is usually drifting with the current and wind. One person wanted Lying to to be added. The extra terms are just there for the convenience of the transcriber. If you see Stopped, Hove to or Lying to you can enter just the first letter instead of having to think about entering an H for Stopped or Lying to.

While it would seem that the ship wouldn't move all that much when it's lying to, I've seen cases where they could be drifting as much as 2 knots, so over night between 2000 and 0800 they've moved as much as 20-30 nautical miles.

[ggordon]

I'm fine with keeping it the same and entering two letters when necessary.

[ggordon]

Missing Characters or Cells Skipped When Entering Weather Data?

The Symptom
When you have rotated the log page image before entering weather data and you enter data too fast, characters will be dropped and/or cells will be skipped.

Don't See the Symptom?
Please send me (ggordon) a message letting me know that you are not seeing a problem when entering data with the image rotated. We are trying to get an idea as to how many transcribers are experiencing this problem.

Possible Fix
We have found that on some computers this can be fixed by changing the graphics options in LibreOffice.

Go to Tools>>Options>>LibreOffice>>View

The Graphics Output options should be in the upper right.

• If Use anti-aliasing is checked, uncheck it.
• If you see the option Use OpenGL for all rendering, check it. This will automatically cause Use hardware acceleration to be unchecked and grayed out.
• If you do not see the option Use OpenGL for all rendering, check the Use hardware acceleration box.

Problem Still Not Fixed?
Send a message to ggordon or michael and we will attempt to guide you through a possible change to your computer system settings that might fix the problem.

Background
Read on if you are interested in knowing more about this problem and the potential fix.

When data is entered in a LibreOffice spreadsheet that has a rotated image in the background, such as our log pages, LibreOffice rotates the image again every time data is entered. This requires a lot of processing power. If you enter data too fast, the computer's CPU is unable to keep up with repeatedly rotating the image and the result is characters getting skipped.

By unchecking Use anti-aliasing, the processing required to rotate the image is reduced. This might slightly reduce the quality of the image, but also allow for fewer characters to be dropped. On the computers we have tried so far, we cannot see any difference in the quality of the image after making this change.

Most computers now include a Graphics Processing Unit (GPU) that is optimized for graphics rendering and can relieve the CPU of the load required for this task. By making the changes listed above you are telling LibreOffice to use the GPU instead of the CPU to rotate the image. However, LibreOffice can only do this if your computer hardware and operating system allow it.

Besides fixing the problem with rotated images, you should also see new log pages load faster, and switching between the weather and events pages go faster.

If you still have a problem after making the above changes, your computer is not configured to allow applications direct use of the GPU or it does not have the necessary hardware to support it. Some computers are set up to require all graphics to be handled by the operating system, which in turn uses the GPU, if available. This causes extra processing to be done by the CPU. Ideally you want to be configured so that the operating system allows LibreOffice to work directly with the GPU. In some cases, just making a change in the Control Panel will allow this. On some computers it's necessary to make a change to the graphics hardware device driver. Ideally you can do a search with your browser to find out what you need to do to allow direct graphics rendering for your specific graphics hardware and operating system. If not, Michael or I will try to help you. However, be aware that not all computers are capable of this. Since your computer hardware is likely different from ours, there is no guarantee that we will be able to help either.

It appears that most computers that have a GPU are by default configured to allow applications direct access to it, and by making the above changes to LibreOffice you will see a big improvement when working with our spreadsheets.

[Michael]

I have been cleaning up the weather data for Kevin, and it is really very clean. I look at each weather column and check for variables out of range, non-numeric characters in columns that should have only numbers etc. etc. I did add a fairly simple verification check to the spreadsheet some time ago. It occurred to me that I could also check the Amount column. Values should be in the range of 1-10. A couple of things I noted: sometimes the data gets dded to the wrong column, so cloud types or clouds from might be in the amount column, the amounts are in the visibility column and the visibility gets put in the Sea state column. This has happened about four times in 100 ship-years. Many ship-years are clean, and some have just one or two errors. Usually it's a case of having two many digits, i.e. 88 instead of 8 etc. Here's an example from one voyage:

1950-03-21 4 SW
1950-03-21 12 SW
1950-06-14 24 E
1950-08-22 13 108
1950-09-13 24 10`
1950-09-24 20 1016
1950-12-13 24 87

My question is this: are you OK if I add the Amount column to the verification module? I don't want to make things to onerous.

PS Interestingly, is that many of the error I pick up are from the typed logs, where the typist has made an error, and the transcriber has done a TWYS, which is correct.

[Randi]

Sounds reasonable to me...

[pommystuart]

I can live with that.

[Hanibal94]

I'm OK with that too.

[ggordon]

I'm fine with adding it. While transcribing earlier today I was thinking it might be good to add verification for more columns. I find the way you have implemented it to be quite unobtrusive.

[Michael]

An Update to Version 3.4 of the spreadsheets has been released.

NEW FEATURES

The Verification module now checks the Clr and Amt columns. The Clr column indicates the amount of sky that is clear. Around 1900 this column was replaced with the Amount of Sky with Cloud.

Bugs fixed

The Hourly Change verification of the Baro column wasn’t dealing properly with a one-hour allowed change. It was converting 0.50 to 1.

Improvements

The Help page has a collection of Tips and Tricks at the bottom. It would be worthwhile to look them over.

---

The updated Version 3_4 spreadsheets have been uploaded to their usual places. Feel free to download them at a time that is convenient for you, but there is no need to do so immediately. Eventually you’ll get this version when you switch to a new ship. About the only things you’ll notice with this update are the verification checks on the Clr and Amt columns, and a few more pressure verification checks for a one hour change being out of range.

Note: All 321 revisions since Version 01-08-C are listed in the Revisions sheet. It makes for an interesting read to see how much these spreadsheets have evolved.

[Michael]

If you wonder how well you're doing, this should be a good example:

When I have completed calculating all the hourly positions, I run several checks on the weather data, looking for values out of range, large changes in successive values, alpha characters in numeric fields etc. I have noticed that transcriptions of typed logs generate many more errors than do transcriptions of ones that are handwritten. The previous voyage I worked on, and which had the original handwritten logs, had just three errors, all from the log keeper, who had two winds SEE and one SWW. This set of typed logs had over 100 errors, I gave up keeping track, and the ones I checked were all transcribed as shown on the log page. I gave up looking at every page flagged, because so many of the errors that I checked were entered as written. Errors that had an obvious fix, like: 29.92 39.92, 29.92, etc, I could fix the error without looking at the page.

So, errors: Log keeper 3, typist 100, transcribers 0.

[Randi]

Go transcribers!

[pommystuart]

"This set of typed logs had over 100 errors,'
I hope that was not my log.

"Errors that had an obvious fix, like: 29.92 39.92, 29.92, etc,'
Maybe mention to the transcriber about Alt +1

Seriously, maybe some of the Newbies do not know about these shortcuts.

[Randi]

It was the original typists that made the errors, not the transcribers.
The transcribers acted correctly and copied what was in the log.

[Michael]

Exactly! The typists needed Alt-1 and automatic verification checks.

Mind you, I sympathize with them. They don't have the image under the data cell, as we do. Neither can they go back and fix anything if they make a mistake. I haven't seen the use of whiteout. I also suspect that it's just a tedious job, and they're not as dedicated as we are. I imagine a group of typists sitting in a room, typing away and occasionally chatting.

When I was on my meteorologist course in Toronto, back in 1971, I went into the room where there were about twenty or thirty ladies key-punching the weather data that had been entered on the daily and monthly weather weather sheets and mailed to our headquarters at the end of each month. It was very noisy with twenty key-punch machines running. One lady would enter the data on a punch card, one for each hour for the hourly sheets, and one for each day on the monthly sheets. She would then pass the cards to another lady, who would put the cards in the machine and she would enter the same data. The machine would compare what was on the card with what the second lady was typing and, if there was a discrepancy, the card would be rejected and it would be re-punched again.

In all the years I looked at climate data, I found only one error. It was in the Old Crow hourly data, and the sequence was something like -34, -34- -35, 36, -36, -34.

Now, with machines taking the weather data there are all sorts of questionable data getting into the records. And, of course, they stop working, so there are lots of gaps in the data.

This is an example of the data on Form 2320. You can imagine how boring it would be sitting there in a noisy room and entering data like this eight hours a day, five days a week.

Code: Select all

Victoria Int'l A                                    Daily Weather Report                                      5 February 1991

Hour    Sky Condition                       Vis Weather          MSLP     TT    Td DD FF   StnP  Clouds                Tw  RH
LST                                          Km                   mb    degC  degC   kmh     mb                      degC   %

0000 9 SCT 40 SCT 100 SCT C250 BKN         24.1                 1011.8/ 10.5/  6.6/20 20/1009.3  CF1TCU1AC3CI3        8.3/ 77
0100 8 SCT 50 SCT 250 -BKN                 24.1                 1013.2/  9.2/  6.3/23 11/1011.0  CF1SC2CI2            7.8/ 82
0200 8 SCT 20 SCT C250 BKN                 24.1                 1015.4/  8.5/  4.7/25 22/1013.0  CF1SC2CI3            6.7/ 77
0300 12 SCT 30 SCT 250 -SCT                24.1                 1017.0/  7.5/  0.6/26 33/1014.7  SC1SC1CI             4.5/ 62
0400 17 SCT                                24.1                 1018.8/  6.8/  0.1/27 22/1016.4  SC1                  4.0/ 62
0500 30 SCT                                24.1                 1020.5/  6.3/ -1.1/26 24/1018.1  SC1                  3.2/ 59
0600 CLR                                   32.2                 1021.9/  5.9/ -1.6/26 20/1019.5                       2.9/ 58
0700 30 SCT 250 -BKN                       32.2                 1023.9/  5.1/ -1.5/26 15/1021.5  SC1CI1               2.5/ 62
0800 30 SCT 250 -SCT                       40.2                 1025.3/  4.3/ -1.5/28 06/1022.8  SC1CI1               2.0/ 66
0900 40 SCT 250 -BKN                       48.3                 1026.5/  5.2/  0.1/27 13/1024.2  SC1CI2               3.1/ 70
1000 250 -SCT                              48.3                 1027.7/  7.3/  0.3/27 07/1025.2  CI                   4.3/ 61
1100 250 -SCT                              48.3                 1028.8/  7.8/  0.0/00 00/1026.6  CI                   4.5/ 58
1200 250 -SCT                              48.3                 1029.0/  8.4/  0.0/09 07/1026.6  CI                   4.8/ 55
1300 250 -SCT                              48.3                 1028.9/  8.8/  0.3/08 06/1026.6  CI                   5.1/ 55
1400 30 SCT 250 -BKN                       48.3                 1028.5/  9.1/  1.5/11 11/1026.2  CF1CI                5.7/ 59
1500 28 SCT 110 SCT 270 -SCT               40.2                 1029.1/  8.8/  1.9/12 11/1026.6  CF1AC1CI1            5.7/ 62
1600 110 SCT 270 -SCT 320 -BKN             32.2                 1029.2/  7.9/  1.2/13 19/1026.9  AC1CI1CI1            5.0/ 63
1700 110 SCT 250 SCT 320 -BKN              32.2                 1029.1/  7.0/  1.6/13 11/1026.6  AC1CI2CI1            4.7/ 69
1800 110 SCT 250 SCT 320 -BKN              32.2                 1028.5/  6.0/  1.9/11 04/1026.2  AC1CI1CI1            4.2/ 75
1900 110 SCT 250 SCT 300 -BKN              32.2                 1028.4/  3.7/  1.1/27 04/1026.2  AC1CI2CI1            2.6/ 83
2000 110 SCT C300 BKN                      32.2                 1028.6/  3.7/  1.6/00 00/1026.2  AC1CI5               2.8/ 86
2100 110 SCT 300 -BKN                      32.2                 1028.0/  3.2/  0.9/00 00/1025.6  AC1CI2               2.3/ 85
2200 110 SCT 300 -BKN                      32.2                 1027.5/  3.0/  1.3/26 04/1025.2  AC1CI1               2.3/ 88
2300 110 SCT 300 -SCT                      32.2                 1027.9/  2.9/  1.0/25 04/1025.6  AC1CI                2.1/ 87


                                                    Daily Weather Summary

Daily Maximum Temperature          10.9 deg C                                   Day with Thunderstorm                 0
Daily Minimum Temp                  2.6 deg C                                   Day with Freezing Rain                0
Daily Mean Temp                     6.8 deg C                                   Day with Hail                         0
Daily Maximum RH                     88 %                                       Day with Fog/Ice Fog                  0
Daily Minimum RH                     53 %                                       Day with Smoke/Haze                   0
6 hr Pcp End 1200Z                  0.0 mm                                      Day with BD/BN                        0
6 hr Pcp End 1800Z                  0.0 mm                                      Day with BS                           0
6 hr Pcp End 0000Z                  0.0 mm                                      Day with Wind >= 28 kt                0
6 hr Pcp End 0600Z                  0.0 mm                                      Day with Wind >= 34 kt                0
Total Rainfall                      0.0 mm                                      FF Xtrm Gust                         54 km/h
Total Snowfall                      0.0 cm                                      Time of Xtrm Gust LST                12 h LST
Total Precipitation                 0.0 mm                                      DD Xtrm Gust 36 pts                 240 degrees
Snow on Ground                        0 cm

[pommystuart]

[Michael]

With the completion of McCulloch 1900, we crossed the 3 million bits of weather data mark.

As of today, the total number of individual weather data in the 108 files with weather data: 3,035,611.
The total number of ship positions is: 864,148

[Randi]