Author Topic: Cloud Code Questions  (Read 10898 times)

okopho

  • Jr. Member
  • **
  • Posts: 54
    • View Profile
Cloud Code Questions
« on: February 02, 2013, 08:46:19 pm »
I second Randi!!  I have had no difficulty so far with it, just think in terms of combining TWYS with KISS. :)

Speaking as a newbie, the issue with cloud codes on the Jeanette caused me quite a lot of hassle. It would have helped a lot if there was a note on the reference page covering this.

After a lot of back and forth editing of previous logs, I settled on the following approach: if the cloud code is clearly a single item on one line, then use the appropriate code from the drop-down list; otherwise, TWYS.

Of course, since multiple lines are being used, it's not always possible to Type Exactly What You See. So I have settled on the convention of using a single space to represent a newline, and relied on the dots and hyphens to maintain clarity. So a three-line cloud code entry might be transcribed like this: cir. cir.strat. strat.

Janet Jaguar

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 10173
  • Smell the sea, feel the sky, & fly into the mystic
    • View Profile
Re: Cloud Code Questions
« Reply #1 on: February 02, 2013, 08:49:41 pm »
I second Randi!!  I have had no difficulty so far with it, just think in terms of combining TWYS with KISS. :)

Speaking as a newbie, the issue with cloud codes on the Jeanette caused me quite a lot of hassle. It would have helped a lot if there was a note on the reference page covering this.

After a lot of back and forth editing of previous logs, I settled on the following approach: if the cloud code is clearly a single item on one line, then use the appropriate code from the drop-down list; otherwise, TWYS.

Of course, since multiple lines are being used, it's not always possible to Type Exactly What You See. So I have settled on the convention of using a single space to represent a newline, and relied on the dots and hyphens to maintain clarity. So a three-line cloud code entry might be transcribed like this: cir. cir.strat. strat.

Perfect.  That will always work for cloud codes. :)

Craig

  • Shipherd
  • Hero Member
  • *****
  • Posts: 3355
    • View Profile
Re: Cloud Code Questions
« Reply #2 on: February 02, 2013, 09:24:40 pm »
I second Randi!!  I have had no difficulty so far with it, just think in terms of combining TWYS with KISS. :)

Speaking as a newbie, the issue with cloud codes on the Jeanette caused me quite a lot of hassle. It would have helped a lot if there was a note on the reference page covering this.

After a lot of back and forth editing of previous logs, I settled on the following approach: if the cloud code is clearly a single item on one line, then use the appropriate code from the drop-down list; otherwise, TWYS.

Of course, since multiple lines are being used, it's not always possible to Type Exactly What You See. So I have settled on the convention of using a single space to represent a newline, and relied on the dots and hyphens to maintain clarity. So a three-line cloud code entry might be transcribed like this: cir. cir.strat. strat.

I am also doing the Jeannette and I am also using spaces to represent a new line. For your example I would have "cir cir-str str".  I always join all components on a line with hyphens. Since the drop-down list always uses 3 character abbreviations I am doing the same, regardless of what is written.  I am wondering how the science team is going to do the comparisons if we don't all use the same convention, though.

okopho

  • Jr. Member
  • **
  • Posts: 54
    • View Profile
Re: Cloud Code Questions
« Reply #3 on: February 02, 2013, 09:47:25 pm »
I am also doing the Jeannette and I am also using spaces to represent a new line. For your example I would have "cir cir-str str".  I always join all components on a line with hyphens. Since the drop-down list always uses 3 character abbreviations I am doing the same, regardless of what is written.I am wondering how the science team is going to do the comparisons if we don't all use the same convention, though.

This is why I thought adding something to the Jeanette reference page might be a good idea. But then again, since many people will never read all the guidance available on the forums, it's not going to eliminate the problem.


Randi

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 13068
    • View Profile
Re: Cloud Code Questions
« Reply #4 on: February 02, 2013, 09:57:19 pm »
...
Speaking as a newbie, the issue with cloud codes on the Jeanette caused me quite a lot of hassle. It would have helped a lot if there was a note on the reference page covering this.
...

You are quite right. The problem is that when we do the Reference pages, we generally have only that specific page. So, we did not know about this problem. The Discussion topics are intended to deal with problems and questions as they come up. I will add a note to the Jeannette's Reference page - or perhaps in TWYS yes but.

Craig

  • Shipherd
  • Hero Member
  • *****
  • Posts: 3355
    • View Profile
Re: Cloud Code Questions
« Reply #5 on: February 02, 2013, 10:13:55 pm »
At least for the Jeannette all of us that are now actively transcribing it are visiting the forum so we can agree on a common method for transcribing the cloud types. The Jeannette is the first ship I have encountered where the cloud codes have been entered so precisely.

If we can at least all agree on using spaces to separate mulitple lines then the science team will have some hope of sorting it out. Also, I see no need to in typing more than three characters of the cloud code. I have been using a hyphen for compound cloud types because that's what the science team uses and I assume that a single line represents only one cloud type (even though it can be compound). If this convention is acceptable perhaps Janet or Randi could send a message to the other transcribers urging them to adopt it?

Janet Jaguar

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 10173
  • Smell the sea, feel the sky, & fly into the mystic
    • View Profile
Re: Cloud Code Questions
« Reply #6 on: February 02, 2013, 10:29:17 pm »
That is perfect.

TWYS in the cloud codes means at minimum that a code for every cloud type or part of a combo cloud type is there.  Precise transcription, or using the drop-down box instead, or combining the two in some way all work.  Do what is comfortable, the climatologists will get usable data from us much more efficiently than if they had to do the transcribing.  Trust them to straighten out antique variations as noted by amateurs.  We are leaving them richer in knowledge than they were before us. ;)

Randi

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 13068
    • View Profile
Re: Cloud Code Questions
« Reply #7 on: February 02, 2013, 10:54:26 pm »
We have removed Jeannette from the ships recommended for new transcribers.
The log format is fairly standard, but the use definitely is not.



Officially, there are two ways to handle cloud codes:
- Type what you see
- Select from the drop-down list

I see Janet beat me to it!

In short, both "cir. cir.strat. strat." and "cir cir-str str" are both correct.
In either case, we will have to resign ourselves to 'space' as representing 'new-line'.
When they are this confusing I would lean towards TWY(think you)S(as best you can), but that is personal preference. If you are confident of your interpretation, you can use the drop-down codes.

okopho - is this a made-up example, or can you give me a link to the page so I can show it?



okopho

  • Jr. Member
  • **
  • Posts: 54
    • View Profile
Re: Cloud Code Questions
« Reply #8 on: February 03, 2013, 12:13:14 am »
okopho - is this a made-up example, or can you give me a link to the page so I can show it?

It's sort of made-up (the dots are a little different) from the 6pm entry in this log:

https://oldweather.s3.amazonaws.com/ow3/final/USS%20Jeannette/vol001of004/vol001_089_0.jpg

Here are a few more logs with a good variety of multi-line cloud codes:

https://oldweather.s3.amazonaws.com/ow3/final/USS%20Jeannette/vol001of004/vol001_073_0.jpg
https://oldweather.s3.amazonaws.com/ow3/final/USS%20Jeannette/vol001of004/vol001_074_0.jpg
https://oldweather.s3.amazonaws.com/ow3/final/USS%20Jeannette/vol001of004/vol001_075_0.jpg
https://oldweather.s3.amazonaws.com/ow3/final/USS%20Jeannette/vol001of004/vol001_076_0.jpg
https://oldweather.s3.amazonaws.com/ow3/final/USS%20Jeannette/vol001of004/vol001_077_0.jpg
https://oldweather.s3.amazonaws.com/ow3/final/USS%20Jeannette/vol001of004/vol001_081_0.jpg
https://oldweather.s3.amazonaws.com/ow3/final/USS%20Jeannette/vol001of004/vol001_082_0.jpg

The writer seems quite consistent in the way he uses multiple lines, and also quite consistetly uses ampersands where two separate items are on one line. However, as you might expect, there are exceptions. For instance, in the 7pm entry in this log:

https://oldweather.s3.amazonaws.com/ow3/final/USS%20Jeannette/vol001of004/vol001_073_0.jpg

he has nimb.cum on one line. Given that he very consistently uses cum.nimb elsewhere, it's hard to know whether it should be one item or two (but, of course, TWYS can deal with that).

He also mostly follows every cloud code part with a dot, and rarely uses hyphens to indicate compound codes. But, then again, there are posssible exceptions. For instance, the 8pm entry in this log:

https://oldweather.s3.amazonaws.com/ow3/final/USS%20Jeannette/vol001of004/vol001_075_0.jpg

consists of two lines, but there seems to be a hyphen between them (which, again, TWYS can deal with).


Craig

  • Shipherd
  • Hero Member
  • *****
  • Posts: 3355
    • View Profile
Re: Cloud Code Questions
« Reply #9 on: February 03, 2013, 02:00:31 am »
That's a good representation of the early part of the trip, okopho. When De Long takes over the log towards the end of 1879 I don't remember any more ampersands or hyphens. Each line is a cloud type, which can be single, double or triple, and a period is used for abbreviations only.

http://oldweather.s3.amazonaws.com/ow3/final/USS%20Jeannette/vol002of004/vol002_129_0.jpg

Here's an example of a triple compound cloud type at 7 PM. http://oldweather.s3.amazonaws.com/ow3/final/USS%20Jeannette/vol002of004/vol002_101_0.jpg

Here's a particularly busy one
http://oldweather.s3.amazonaws.com/ow3/final/USS%20Jeannette/vol002of004/vol002_023_0.jpg


Randi

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 13068
    • View Profile
Re: Cloud Code Questions
« Reply #10 on: February 03, 2013, 07:42:55 am »
You have probably seen it, but I just wanted to remind you of an early comment from the science team

We knew these logs were going to be quite a bit harder than the ones we're used to, and we've had some serious debates in the science team about whether it was worth the effort. But we are in no doubt - these are the logs we need to be looking at - we're not inflicting all this trouble on you for no reason.

If you've got 48 seconds to spare, watch the video at http://vimeo.com/43723075. It's a bit frenetic (and it goes backwards in time) but it does show the observations that we've currently got - that's what we're using to understand how the climate varies and changes over decades and centuries. If you look at the North Pacific, you'll see that in the nineteenth and early twentieth centuries there are almost no observations, and that's a big problem - Pacific climate variability matters to a lot of people. That's why Kevin is currently shivering up in the Chukchi sea and it's why we're asking you to battle with these more challenging logs. We need these observations even more than we needed the RN WW1 ones.

When we did the Royal Navy logs we knew what they all looked like when we started building the website. The US logs are older, more variable, and we're finding the logs and building the website at the same time. This makes Arfon and Stuart's job much harder, and that makes it harder for us all. It will get better - but it's always going to be a different challenge from the old version.

So there's a strong science payoff for all the hassle, and almost certainly some fascinating new stories and history - certainly I'm expecting more strange and unexpected events to occur in these logs. Remember also that it's not a race - relax: if the site's misbehaving, knock off for a tot of rum (or a Devon cream tea), it'll still be there when you return. And rest assured that we hear what you're saying and we're trying hard to put up lots of really interesting and valuable logs, and to knock the bugs out of the user interface.

Philip

Randi

  • Global Moderator
  • Hero Member
  • *****
  • Posts: 13068
    • View Profile
Re: Cloud Code Questions
« Reply #11 on: February 03, 2013, 07:52:51 am »
okopho - Thank you for the links.
Yes, I think this is one where it is quicker to TWYS than to try to convert it to drop-down equivalents ;)

Craig

  • Shipherd
  • Hero Member
  • *****
  • Posts: 3355
    • View Profile
Re: Cloud Code Questions
« Reply #12 on: February 03, 2013, 11:12:49 am »
okopho - Thank you for the links.
Yes, I think this is one where it is quicker to TWYS than to try to convert it to drop-down equivalents ;)

That may work best for the early part of the voyage but once you get to De Long's pages you will find that you can't rely on periods to indicate a compound cloud code. If you TWYS then you will put spaces between the components and these will no longer indicate only a line feed. (see my examples).


okopho

  • Jr. Member
  • **
  • Posts: 54
    • View Profile
Re: Cloud Code Questions
« Reply #13 on: February 03, 2013, 07:45:12 pm »
If you TWYS then you will put spaces between the components and these will no longer indicate only a line feed. (see my examples).

ISTM that the most important thing is that cloud codes entered via the drop-down list are clearly distinguishable from the ones transcribed directly. That way, the science team will always know which cases they need to treat differently. So the precise details of the format used when directly transcribing the cloud codes should not matter too much.

Currently, the pre-defined items in the drop-down list only contain a fixed range of alphabetic characters, plus one hyphen. So any transcription format that includes at least one character not in that set (such as a space, dot, ampersand, etc) should be enough to distinguish it from a pre-defined cloud code.

The real issue is that most users encountering this for the first time:

Quote
cum
nimb

might naturally assume that the writer meant Cumulonimbus and wrongly select Cum-Nim from the drop-down list. So some prominent warnings regarding multiple cloud codes could be very helpful.

Pommy Stuart

  • Shipherd
  • Hero Member
  • *****
  • Posts: 3609
  • A closed mouth gathers no foot.
    • View Profile
Re: Cloud Code Questions
« Reply #14 on: February 03, 2013, 09:14:47 pm »
This is turning into an interesting discussion.
Maybe it is time for a definitive answer from the Boffins who will be getting our data so that they can understand what we are giving them.
I  have not done the Jeanette but I have been entering the like of

cum
nimb

as cum-nim from the drop down list, based on the fact that there was not enough room for the scribe to run them together across the column. I will continue to do this till told by the Boffins to change.

My question to the Boffins is, what effect will it have on their modeling if the clouds are recorded as Cum Nim (possibly indicating a change within that time period) or Cum-Nim (indicating Cumulonimbus) when you usually have from 8 to 24 data entries for each day?
« Last Edit: February 03, 2013, 09:49:26 pm by Pommy Stuart »