
Tips and guidelines for getting help with NCL and posting to email lists
This is a page of tips for getting help with NCL and guidelines for posting to the NCL email lists (ncl-talk and ncl-install).
Tips for getting help with NCL
There are hundreds of users on these lists. The large majority of NCL email subscribers are unaffiliated with NCAR and are not reimbursed in any way for responding to NCL questions. They are, just like you, busy researchers who are happy to help out whenever they can.
To avoid cluttering the archives with commonly-repeated questions or unnecessarily filling up email inboxes, please first read over these tips for getting help with NCL questions. If these tips don't help, then read the posting guidelines below.
- Do you want to report a bug?
Submit an "issue" via our NCL GitHub repo. Go to the "Issues" tab and click on the "New issue" button.
- Are you completely new to NCL?
A good place to start is with the "Getting Started With NCL" page.
- Are you getting an error message you don't understand?
Read the "Error messages" section of our FAQ, or read through this list of common error messages, what causes them, and how to fix them.
- Read your error messages carefully.
Some can be very useful, some can be pretty obscure admittedly, but at the very least you should find the line of code where the problem occurs and a message that you can use to search the ncl-talk archives.
- Try a quick search of the ncl-talk or ncl-install archives.
For example, if you want to know how to calculate wind speed from WRF data, enter "WRF wind speed" in the ncl-talk search box.
- Is the question about a function or procedure?
For example, questions like "is there a function for calculating the average of my data" and "how do I use NCL's 'avg' function?" could both be answered by looking at the NCL function list and the related function documentation page for avg.
- Is the question about analyzing or plotting WRF data?
See the WRF examples page or send email to wrfhelp@ucar.edu. You may also want to check out our Python package, wrf-python.
- Look at the NCL examples page.
This page contains a wide variety of topics with hundreds of examples.
- Look at the
"tips" page.
If you have a question like "how do I put a time stamp on a plot" or "how do I create some dummy data", or "how do I interpolate data from a HOMME grid to a POP grid", then on this page, use your browser's "search" to search for keywords like "time stamp", "dummy", or "HOMME".
- Look at the Frequently Asked Questions (FAQ).
- Look at your data before assuming that NCL is the problem.
- Use the printVarSummary procedure to examine
your variable's dimension sizes, type, attributes, _FillValue attribute
(if any), coordinate arrays, etc.
- NCL does not recognize "missing_value" as a valid missing value
attribute; only "_FillValue" is recognized. If you have a "missing_value"
attribute and not a "_FillValue" attribute, then you need to do the
following before you do any calculations on or plotting of your data:
x@_FillValue = x@missing_value
See this write-up on missing values for more information. - Use the printMinMax
procedure to verify that your data values are in the correct range.
- Check if there's a scale factor and/or an offset attribute that
needs to be applied to the data. These will usually appear as
attributes attached to your variable, like "scale_factor" and
"add_offset". If your data is of type "byte" or "short", then
sometimes this is an indication that you need to apply a scale factor
and/or an offset. See
the byte2flt
and short2flt functions for more
information.
- Some computational and plotting functions require your data be in the correct units in order for them to work properly. Check the appropriate function page for more information.
- Use the printVarSummary procedure to examine
your variable's dimension sizes, type, attributes, _FillValue attribute
(if any), coordinate arrays, etc.
Guidelines for posting to ncl-talk or ncl-install
If the tips section above didn't help, then please use these guidelines for posting a question to one of our email lists:
- Is the question about NCL?
The email lists are for NCL-specific questions, so avoid general questions about data analysis, statistics, meteorology, etc.
- Is the question about installing NCL or setting up your
environment to run NCL?
If so, post the question to ncl-install and not to ncl-talk.
- Use an informative subject and not just "help" or "request"
or "my script doesn't work".
- Do not start off your email by addressing a single person
directly, like "Dear Mary" or "Hello Dennis". This may cause other
people to ignore the email. This is a email list of hundreds of
users, so address the whole group.
- If you have further questions about an answer you received on
ncl-talk, please include ncl-talk on your follow-up email,
unless the responder has purposely taken the conversation offline.
- If somebody responds to one of your ncl-talk questions, and you
have another question to ask, please do not email the respondee
directly. All NCL questions should be emailed to ncl-talk with
a new subject line.
- Be as detailed as possible. If you have a question about a script
that is not working, then include the following information where
appropriate and/or possible:
- The version of NCL you are running (ncl -V).
- The exact error message, if any.
- A clean NCL script that illustrates the problem or the
question you have. Short and clean scripts are even better!
Please make sure you send us a script we can run without any
modifications. Sending us partial code or code that requires editing
will likely delay us in getting an answer to you.
- Any data files needed to run your script. You can use
our ftp account to upload
large (> 1.5 megabyte) files
- An image (if appropriate) (PNG, PS, or PDF file) that illustrates
the problem.
- The type of system you're on (usually "uname -a" is enough).
- Do not attach files larger than 1.5 megabytes. If you need to
give us a large file, put it on our
ftp account.
- The version of NCL you are running (ncl -V).