New software, same as the old software

DKNewsham

In my department, we have several NMR spectrometers from a single manufacturer. The data from those instruments is collected and processed using IRIX machines connected to the instruments. There is a GUI, but it it's counterintuitive and most users simply memorize a series of commands to type that will collect data. For example, a typical experiment can be performed by typing: edc, rpar, wobb, lock, eda, ii, rga, zg, lb, efp, apk, abs. Most of those commands bring up a dialogue window where setting like ns and prosol can be edited. Then the user gets to use the GUI to actually interpret the data. However, despite the terrible UI and memory leaks, it works and produces good data.

Recently, one of the machines was updated to a RedHat box with the latest version of the software from the manufacturer. It somehow maintains all of the quirks of the old software while being less functional and producing lower quality reports. As far as I can tell, it's still the old software with a layer of python to make the modern UI work. Here are just a few of the new problems. I'm sure there are plenty more to be found.

There are two buttons that can be interpreted as stop. The square is "halt" which ends data acquisition and retains the data collected so far. The octagonal stop sign is "stop" which ends data acquisition and discards the data collected so far. Don't mix those up. Also, the buttons to the left are rather unkind to the colorblind users (of which there are several)
This informative dialogue comes up if when you try to clear the automatically selected integral regions from the spectrum.
After collecting a spectrum from the instrument, the user's data is stored in /opt/SOFTWARE_NAME/data/USER_NAME/nmr/DATA_SET_NAME/ Did I miss the memo where everyone agreed to start storing user data in /opt? Also, there isn't actually a list of users. Everyone is logged in as a single user and can type any arbitrary username they like. This leads to mountains of users with name, Name, NAME, Namf and so on.

ounos

Don't hold it back. I'm sure there is more of it. Good stuff.

NSCoder

[url=http://www.science-groove.org/Now/]Hooray for NMR Spectroscopy![/url]

Is that a penguin between the shopping list and the bar chart?

And why do you need a button to type a carriage return?

WTF

@Zecc said:

Is that a penguin between the shopping list and the bar chart?

Don't be silly, it's the famous Canadian Blue Xmas tree!

DKNewsham

Sorry, but there is nothing as cool as a penguins or Christmas trees in this monster. That button activates the tool to measure peak widths. The carriage return button is "return" which exits the current mode without saving changes. It's brother, carriage return with floppy disk, saves and returns.

Here is a larger screenshot showing just how complicated this app is for new users. You can see the CLI at the bottom of the screen. This shot is from integration mode (Yes, that highlighted icon is supposed to be an integral sign). Some buttons appear twice in different toolbars, but perform the same function. The colored squares and triangles in the upper-right represent different data sets. The developer felt the need to invent a new UI element even though they used tabs elsewhere.

There are many other apps that manage to accomplish the same tasks in a much simpler way. Most of them even have anti-aliasing. I usually end up acquiring the data and then processing it on my own machine with one of those other apps. Of course, they don't cost thousands of dollars per seat so they must be missing some of the awesome features of this app, not that I could find them in this mess.

Just in case anyone cares, the spectrum shows diethyl ether, pyridine and methyl dichloridophosphite.

Nandurius

Fabulous story, and awful application. Why bother hiding the name of the program though? There aren't exactly a lot of tools on the market that have commands such as edc, rpar or wobb. Make the name of the application rather easy to find out.

DKNewsham

@Nandurius said:

Fabulous story, and awful application. Why bother hiding the name of the program though? There aren't exactly a lot of tools on the market that have commands such as edc, rpar or wobb. Make the name of the application rather easy to find out.

I'm fairly new to this forum and I had noticed that the name of the company, person or software of interest is often changed or omitted. I figured that this is either a professional courtesy or a flimsy defense against angry lawyers. In any case, I went with the flow and avoided any mention of the application or its developer. There are really only a handful of NMR manufacturers so it would have been impossible for me to stop anyone with access to Google from figuring it out in a few minutes.