Pushing Excel to the limits. Matlab? HA!



  •  In one of my postgrad courses (it was about electronic instrumentation) I had to design and simulate a simple 6-pole Chebychev filter cascade as the second homework assignment.

    Since :

    • I was too lazy to dig out PSpice or LabView from boxes of my old pregraduate-years CDs.
    • A filter design program I had couldn't design filters with a particular topology..
    • I sure as hell wouldn't go through a 1-hour MATLAB installation just to use the built-in chebychev design, complex number support and function plotting functions...

    .. I just decided to make an ad-hoc solution using Excel.

     The best of all is that it actually worked. Judge for yourself:

     (max 10 times download limit, d'oh!)

     

     



  • On my previous job I was using Excel to simulate and tabulate various aspects of an CMOS sensor image processing pipeline, like color correction and such. The best part of it: you could have a live update of any tweak. You could drag points on a graph and see what the result. And other things. Other people were using Matlab, but I could not be bothered to. My former boss was using Excel with a bit of Basic in it to simulate auto-exposure behavior.



  •  This isn't quite as bad, but my last summer job was working at a plant, where I had to create a whole system of Excel spreadsheets for making daily and monthly reports - there was a daily template that got copied for each day, pulled in numbers from a CSV file with a shitload of data (minute-by-minute for ~80 sensors), and export the averages to a text file and upload to a central system (by writing FTP commands to a file then doing an "ftp.exe < commands > output" and parsing the output to find errors since the return value for the FTP program seems to have absolutely no meaning).  The monthly sheet then had a macro which did a find-and-replace to set references to the individual daily sheets, and some more stuff with macros too horrible to mention.  I got to know VBA MUCH, MUCH better than I ever wanted to.

    Believe me, I would've loved to generate it all with a bit of Python or something but Excel was the only thing available (I tried Access reports but those sucked too).



  • @burntfuse said:

    .. horrible story about abusing data ..

    I'm curious to know how the CSV data was created in the first place. Logging data from plant equipment, processing it and generating reports is one of my work areas, and there are a large number of packages that will do this sort of thing for you. It sounds like the owners either cheaped out on buying the correct software or didn't know what it could do.



  • @C4I_Officer said:

     (max 10 times download limit, d'oh!)

     

     

    Not anymore ;-) Unlimited (I think) downloads from here. The real WTF now would be if anyone else besides me finds this useful. In case you do...well...I have a high-pass version of this "filter" too ;-)

    http://www.verzend.be/download/2686/ChebyLowPass-xls.html



  • TRWTF is that people use Crapidshare.



  • I'm using drop.io until they run out of money.

    drop.io ftw



  • @OzPeter said:

    I'm curious to know how the CSV data was created in the first place. Logging data from plant equipment, processing it and generating reports is one of my work areas, and there are a large number of packages that will do this sort of thing for you. It sounds like the owners either cheaped out on buying the correct software or didn't know what it could do.
     

    There was a function in the internal scripting language of the program used for the GUI that exported a bunch of logged data to a CSV file.  There was actually a database program (built on MS SQL maybe?) that would grab all the logged data and store it, so it could be queried, reports could be run, etc.  I know the first thing I thought when I saw that is "why don't we use it?", but there was some reason not to that I don't quite remember (maybe because no programmers were there full-time to deal with the logging, and the other guys only knew Excel, so they wouldn't have been able to figure out how to add new daily averages and stuff when new equipment was put in or something).

     Actually, the internal scripting language is something else I could rant about (strings are limited to 131 characters, no arrays, all IFs are nested, loops exist but are a bad idea in practice since they locked up the whole program and kept data from being collected while they were running, same situation I think with functions...).



  • Mathematica is way better than MATLAB.  At least, that's what Stephen Wolfram told me while clubbing me over the head repeatedly with a copy of A New Kind of Science.


Log in to reply
 

Looks like your connection to What the Daily WTF? was lost, please wait while we try to reconnect.