The last few weeks have been rather frustrating. My task was something that sounded relatively simple, and I’d done it before. Though this time everything took three times as long… Take a glimpse into the life of a PhD student!
So this was the big idea: Make model run with new input data.
The tasks were as follows:
1. Finalise the model domain
2. Download new data
3. Upload new data to the big computer
4. Go through the three steps required to run model
5. Look at plots of output and feel smug.
What actually happened:
- Domain finalised, a nice picture made and I used it in a talk.
- Attempt to download data. Can’t work out where it is on a maze-like website. Get friend to help. He has to find his download scripts. Locate download scripts, they don’t work. Try multiple times. Find scripts that work. Download. Takes ages. Watch some Mitchell and Webb on youtube while we wait.
- Upload data to big computer…. sit and wait.
- Discover data is in wrong format. Repeat steps 2 and 3.
4.1. Data now hopefully in right format. Attempt to put it into model. Discover data is in two different formats. Email help dude. Go back to step 2 and download surface and atmosphere separately.
4.2. Data now separate. Confirm its in two different formats. Put the atmosphere data in! Hurrah! Discover the surface data, which is in an older format, doesn’t work.
4.3. Attempt to convert surface data into newer format. Have to do this on colleagues laptop as the program won’t install on mine. It doesn’t work. It doesn’t recognise surface temperature. Email help dude.
4.4. Go back to step 2 and download slightly different data. It still doesn’t work. Repeat.Call a halt. Go back to step 2. Email help dude again.
2. Download different data from the usual place.
3. Skip this step as I could download straight to the big computer.
4. Go through the three steps of model running. Have minor issues but get it to run!
5. Model still running!
UGH! Such a stress! All this faff with the alternative data that is possibly slightly better for the polar regions has taken two to three weeks (and all summer if you count the time it took to get authorised to download the data in the first place!) However to get it to go with the ‘normal’ data took less than an hour. But its all part of a PhD isn’t it? There is nothing I can do about it! I must say I couldn’t have even tried without the help of my friend, so I’ll be thankful for that.
Tomorrow I should have output, I should be able to look at pretty plots, then I can move to the next phase of work. To more frustration and more going round and round in circles no doubt!