Hah, I like that one. The other classic is Right Click -> New -> Shortcut -> cmd.exe in an explorer "open" window, typically one in an otherwise very locked-down environment.
This has recently got me service access on an old (but new in 2009!) ultrasound machine, for example, for getting raw data and dicom images off in a hurry when the proper authentication details were lost...
The real boss move was navigating a machine with a UI that involved a trackball, keyboard, touch screen(s), touch pad, weird array of custom buttons and a truely stupid menu system.
Configuring US machines is horrible.
But my major US machine rant is them burning metadata into the images (rather than displaying DICOM tags as an overlay). It’s is beyond ridiculous.
Exactly! MR ("my" modality) has it right -- raw data and reconned images are very, very different and although most raw data never ends up in a dicom the mere fact that you genuinely could reconstruct dramatically different bits of info (e.g. magnitude vs phase images) means that the vast expanse of the dicom spec is wide enough to encompass all possible metadata requirements.
US machines do a lot of fun physics on proprietary FPGAs. For inexplicable reasons, every one I've ever worked with or done echo with saves the images as some variation on a theme of screenshots, shoehorned badly into a dicom wrapper, with the metadata burned at 640x480 px (or similar) on top. Even for clever derived modes like doppler -- even for annotations showing things like cardiac E/E' or E/A. They are laptops with a custom pcmcia / pcie card and a 100k-UNIT_OF_CURRENCY price tag, inevitably running a shitty OS with a shittier custom UI...
MRI is my modality of choice too. I’m currently loving most of what Siemens is up to (with some notable exceptions).
The hell of US knows no bounds. Most modalities calibrate a display and then display images (with varying degrees of post processing). US calibrates the screen, sometimes with each boot or even each probe change. Their black levels are abysmal.
> saves the images as some variation on a theme of screenshots
GE has a habit of making DICOMs from screen grabs. I’ve seen it on their PET, CT and MR systems. It causes irritating problems - like reference lines won’t work so you can’t cross reference.
This has recently got me service access on an old (but new in 2009!) ultrasound machine, for example, for getting raw data and dicom images off in a hurry when the proper authentication details were lost...