Reader reactions: Excel in the limelight

Week in Review now includes more of you.

Read the whole story

 

[Personne]

Ego notwithstanding, he knows how much his work is appreciated.

 

[koolraap]

I like this style of wrap up. Good work.

 

[Trout Mask Replica]

Now if we can only get people trained in the use of other software we can be freed from "I only know Excel, therefore, I shall view the world through it." Excel is like the Matrix for the vast majority of office workers. They know nothing else. Need a database? Excel. Need to provide functional specs for your developers? Excel. Need to write a proposal? ( yes, I've seen an entire 15000+ word 'document' in Excel ) Excel! Need to bake a loaf of bread? Say it with me, now! Excel! Excel! Excel!

The mind of the office drone is devoid of most aspects of humanity, as is desired by the corporate soul-suck machine, but when it comes to finding ways to use Excel for which even the craziest lunatic at Microsoft could not even dream of, these people will find it. There is no perversion too taboo, no need so intricate it cannot be twisted to fit within Excel's work sheets, no Satanic rite too foul... Excel is the One True Way for All.

 

[ChrisSD]

Trout Mask Replica[/url]":26kbmgim]Now if we can only get people trained in the use of other software we can be freed from "I only know Excel, therefore, I shall view the world through it."

Oh god that annoys the heck out of me at the moment. The new boss has decided we should throw away our current tools and use Excel for everything because he knows Excel. I'm now trying to explain to him why this is madness.

 

[Deleted member 192806]

Trout Mask Replica[/url]":1o5b689f]Now if we can only get people trained in the use of other software we can be freed from "I only know Excel, therefore, I shall view the world through it." Excel is like the Matrix for the vast majority of office workers. They know nothing else. Need a database? Excel. Need to provide functional specs for your developers? Excel. Need to write a proposal? ( yes, I've seen an entire 15000+ word 'document' in Excel ) Excel! Need to bake a loaf of bread? Say it with me, now! Excel! Excel! Excel!

The mind of the office drone is devoid of most aspects of humanity, as is desired by the corporate soul-suck machine, but when it comes to finding ways to use Excel for which even the craziest lunatic at Microsoft could not even dream of, these people will find it. There is no perversion too taboo, no need so intricate it cannot be twisted to fit within Excel's work sheets, no Satanic rite too foul... Excel is the One True Way for All.

They can even cause you to lose your job.

 

[readykilowatt]

Trout Mask Replica[/url]":2esoucby]The mind of the office drone is devoid of most aspects of humanity, as is desired by the corporate soul-suck machine, but when it comes to finding ways to use Excel for which even the craziest lunatic at Microsoft could not even dream of, these people will find it. There is no perversion too taboo, no need so intricate it cannot be twisted to fit within Excel's work sheets, no Satanic rite too foul... Excel is the One True Way for All.

What's worse is Excel teamed up with sharepoint. Now, not only do you get a list of whatever in Excel, you get a link over Email that is impossible to use because Microsoft didn't bother with a proper naming system when they puked out Sharepoint. At least in the SMB days you could see a path and folders. Now it's a bunch of http://teamcompany.com/%first%quater%sales%figures/bob%collins/phil/phil's%first%quarter%sales%report.xls links in email, which, when clicked on, sends you to the excel viewer, an app for only the most basic spreadsheets, forcing you to click on the "open in Excel" button, which (thanks to the IT department's lack of faith in us), gives us a continuing pop-up warning about downloading files off "the Internet" and then asks if we want to open to edit, or to read-only. Much easier than just pointing to a drive, right?

 

[rimbaud]

Best. Roundup. Ever.

More of this please

 

[_fluffy]

I like how Apple's Numbers, for all its problems and shortcomings, will visually SHOW YOU what ranges of numbers an expression refers to, if you highlight the expression's cell. It's a great little double-check tool that I use all the time, since mostly I use Numbers for doing balance sheets for my personal budget and home business and such.

 

[firsttimeposter]

_fluffy[/url]":2ubuix1k]I like how Apple's Numbers, for all its problems and shortcomings, will visually SHOW YOU what ranges of numbers an expression refers to, if you highlight the expression's cell. It's a great little double-check tool that I use all the time, since mostly I use Numbers for doing balance sheets for my personal budget and home business and such.

Excel shows you the ranges as well. I don't know why everyone is bashing excel, it's just a tool. I am not going to say c++ is shit because someone makes a buggy software with it.

 

[pokrface]

Aurich's image-production prowess is peerless. I'm pretty sure he's actually a wizard.

 

[Grieviant]

The two most troubling things about the error in the economics paper:

(1) It wasn't caught until 3 years after the fact, the damage of the paper being used to develop austerity recommendations that influence millions of people already being done. Then, once people started poking further into some of the assumptions and questionable exclusions of available data sets in the paper, more questions were raised as to the validity of its results. Meanwhile, the authors steadfastly maintain that none of this has a serious impact on its general conclusions, as if the numbers don't mean a thing (0.5% reduction, 2% growth ... what's the difference?). The authors were even gracious enough to say "we're not mad at the other researchers who found our mistake". Give me a break.

(2) That highly paid researchers actually use a program like Excel for scientific computing. Perhaps I'm simply ignorant of the accepted 'tools of the trade' for Economics work? There are myriad math and stats packages that seem better suited to the job. Using the right tools does not eliminate the chance of mistakes, but it certainly reduces it.

 

[mrsilver]

Maybe it's just me but I've found Excel (2003 and onwards) pretty good when it comes to warning you that cells may have been left out of a range.

Sure, it isn't perfect - but it's far better than not saying a thing.

 

[Big Wooly Mammoth]

mrsilver[/url]":2yvp5c19]Maybe it's just me but I've found Excel (2003 and onwards) pretty good when it comes to warning you that cells may have been left out of a range.

Sure, it isn't perfect - but it's far better than not saying a thing.

This is all true but the comment above yours by Grieviant sums up my feelings. Why the hell would Excel be used in this fashion when there are plenty of stats applications out there (SAS, SPSS, etc.) that are actually designed to evaluate large data sets.

 

[rockforbrains]

Grieviant[/url]":31wk6yai]The two most troubling things about the error in the economics paper:


(2) That highly paid researchers actually use a program like Excel for scientific computing. Perhaps I'm simply ignorant of the accepted 'tools of the trade' for Economics work? There are myriad math and stats packages that seem better suited to the job. Using the right tools does not eliminate the chance of mistakes, but it certainly reduces it.

Spreadsheets are fairly easy to learn to use a modest level. They are excellent for light duty analysis but not really good for truly serious work because the formulas are very difficult to debug and maintain. But many try to use in ways they are not designed. I am surprised these two "economists" were not using more sophisticated tools or even writing their own program. If the data can be crunched on Excel I would expect about common computer language including the scripting languages could handle the number crunching.

 

[crm-114]

_fluffy[/url]":12lgzgpd]I like how Apple's Numbers, for all its problems and shortcomings, will visually SHOW YOU what ranges of numbers an expression refers to, if you highlight the expression's cell. It's a great little double-check tool that I use all the time, since mostly I use Numbers for doing balance sheets for my personal budget and home business and such.

Excel does this as well you just have to opt in. It's called Trace precedents and dependents.

http://www.owenbloggers.com/2011/08/19/ ... recedents/

That said Excel does still have some issues. One is conditional formatting of duplicate values. I've seen it highlight supposed duplicates that were in fact unique.

 

[MrFred]

Grieviant[/url]":12nzzcyn]The two most troubling things about the error in the economics paper:

(1) It wasn't caught until 3 years after the fact, the damage of the paper being used to develop austerity recommendations that influence millions of people already being done. Then, once people started poking further into some of the assumptions and questionable exclusions of available data sets in the paper, more questions were raised as to the validity of its results. Meanwhile, the authors steadfastly maintain that none of this has a serious impact on its general conclusions, as if the numbers don't mean a thing (0.5% reduction, 2% growth ... what's the difference?). The authors were even gracious enough to say "we're not mad at the other researchers who found our mistake". Give me a break.

The paper wasn't peer-reviewed and the data wasn't really available. The data used to attack the paper was provided to the authors directly by Reinhart-Rogoff so that they could study it - it wasn't easily available. Source: Researchers Finally Replicated Reinhart-Rogoff, and There Are Serious Problems

Oh, and an academic counter to it came out six months later, as cited by Herndon et al.. I think the academics wrote off the paper early and moved on.

From what I'd read of it, this seemed more a case of non-academics, i.e. pundits and politicians, grabbing something they liked and running with it. The paper itself supposedly doesn't try to to show causality, which is a problem since it seems reasonable to assume that a country in economic trouble might run up some debt too. Krugman's been going at it all week on his blog.

I agree, though, that the academics should have gone after it once it got picked up by the public -- demanding the spreadsheets, etc. -- if for no reason than professional pride since people who hear about this just end up trusting academics less, ignoring the pundits and politicians' role in this. Part of the problem is that the general public is perfectly fine with the highly filtered news they get, and isn't willing to really get into details.

Grieviant[/url]":12nzzcyn](2) That highly paid researchers actually use a program like Excel for scientific computing. Perhaps I'm simply ignorant of the accepted 'tools of the trade' for Economics work? There are myriad math and stats packages that seem better suited to the job. Using the right tools does not eliminate the chance of mistakes, but it certainly reduces it.

If they'd been by convention compelled to require the data with the paper, even in excel format, this would have been quashed back in 2010, instead of academics going "Well, my data" is better than your data" and not really getting anything accomplished. The data, and the decisions in how to treat the data are part of the argument, indeed the most important part in papers like these, and it's silly to not include it when it is the heart of the argument. It's trivial to include files on journal websites.

I think not normally releasing data is the heart of this problem but the coding error was probably just a case where the problem has grown a little beyond what a spreadsheet is good for -- simple plotting and maybe a little data analysis.

 

[Cartigan]

firsttimeposter[/url]":27qp4pxu]

_fluffy[/url]":27qp4pxu]I like how Apple's Numbers, for all its problems and shortcomings, will visually SHOW YOU what ranges of numbers an expression refers to, if you highlight the expression's cell. It's a great little double-check tool that I use all the time, since mostly I use Numbers for doing balance sheets for my personal budget and home business and such.

Excel shows you the ranges as well. I don't know why everyone is bashing excel, it's just a tool.

The problem with Excel is it is a hammer and office workers look at everything like its a nail.
As a developer, I can't tell you how much I would like to set fire to every piece of standalone software that is instead developed as some sort of Microsoft Excel-dependent faux-program

 

[ColBatGuano]

firsttimeposter[/url]":1t8jqz2x]

Excel shows you the ranges as well. I don't know why everyone is bashing excel, it's just a tool. I am not going to say c++ is shit because someone makes a buggy software with it.

C has a poorly designed syntax - the number of errors that come from =, == mistakes is legendary. This mistake has been inherited by many other languages as well.

Excel has a similar problem - you can do good things with it, but doing so is harder than it should be because they have ignored the fact that people make mistakes and designed the system to prevent that (a harm minimisation strategy).
If people continually make the same mistakes over and over again in a way that a better designed tool would prevent, isn't it time to blame the tool?

 

[firsttimeposter]

I will happily accept excels fault when people will also start blaming other tools and programming languages for the developers mistakes. However you wont see ars calling out C++ because someone somewhere wrote a buggy program or a program which gives incorrect results.

 

[Aurich]

Great job Megan, I love seeing the stories around the stories. Appreciate the kind words too everyone, it's a nice break from my more "responsible" work to bang out little 5 minute photoshops for our writers. My favorites are often the ones where they throw out a ridiculous challenge to me.

 

[Postulator]

I agree with the concept that "Excel is single user". I have had to decipher and repair others' spreadsheets many times, and there are just so many simple errors that can be made, that grow into large numbers very fast.

I am an accountant by trade, and contract to large organisations. I get to see where people have updated formulae and left off the final column/row. People have used the same spreadsheet for years and now nobody knows how it is supposed to work.

In other cases, an Excel "expert" comes in and creates the ultimate Excel tool - using macros, buttons and thingamajigs, before leaving the organisation which has to maintain this "tool" that has no user documentation or maintenance manual. Yes you can program in Excel - it is generally a very bad (TM) idea.

Excel does have "audit" tools to pick up errors in spreadsheets, but these only identify some of the more obvious errors, and even then can make mistakes of their own. Although "trace precedents" and "trace dependants" are invaluable when figuring out someone else's spreadsheet.

 

[HiredGoons]

True on so many levels. Excel's great for manipulating large sets of data, but not if you have to share the files with more than one person.

I was temping with an equities firm for awhile, and even though I'm not exactly a database pro, I kept trying to tell my manager that the sort of getup he wanted done really should have been taken care of with a database tool. Having something that multiple people can freely edit that can refresh itself from external sources is not really something Excel is good at doing.

But . . . he didn't want to spend any money on the software. I'd wager a lot of decisions that involve using Excel over something more appropriate boil down to money.

 

[jiffle]

So often Excel is used as a quick-fix... that then becomes permanent.

Investment Banks seem to be continuously battling millions (literally!) of little spreadsheets that contain various model calculations that are hidden from the management. (I know of at least one company who has made good money creating tools to audit this.)

Excel is, however, a useful indicator: whenever people start creating lot of little spreadsheets to manage day-to-day tasks it's a sign that the software that they are expected to use is flawed, problematic, or not understood by its users. Or, of course, it could be a sign that there is not software filling that role at all, and something needs to be deployed. Either way, extensive use of Excel is a sign of trouble.

Finally, I've always viewed Excel as a poor solution for what most people wanted to do: Take numbers and summarise/report it in meaningful ways.

Almost twenty years ago, Lotus Improv showed that there were better, simpler ways of doing these things, and while Excel has absorbed many of the concepts that Improv pioneered, it has done so at the cost of ever increasing complexity - something that Improv managed to avoid.

 

[ikhtiar]

I am surprised to see many people suggesting to write programs to find the average of 12 numbers. They must find it easier to code than to write a simple formula in Excel. But for most people, it is very similar to use a cannon to kill a mosquito. There are sophisticated programs which can do much more but in my view, Excel does 99% of the tasks with 10% of the efforts. I am not saying SPSS or other tools are useless but if you consider the learning curve, Excel is much better for simple tasks like finding averages. Just because you are an Economist does not mean you always have to use the most advanced but considerably more complex tools.

Before you jump on me, please note that the problem mentioned here is clearly an user error. No software will ever be able to give correct output if the input is wrong. There is a specific term in programming - "Garbage In, Garbage Out". The reaction from the researchers in question convinced me that they did not try hard to check for errors because the erroneous formula generated the result they wanted, why bother checking.

Frankly, all you need to do is to press F2 in Excel to see which cells are included in any formula. I am sure it is much easier than to write a separate program.

On top of that, recent versions of Excel provide visual indications when you omit adjacent cells from a formula. This is precisely what happened and just a "look/glance" would have spotted the problem. But if you are not going to look for it, you will never find it. It would have not matter which software you are using.

 

[Cartigan]

firsttimeposter[/url]":1drvcqiq]I will happily accept excels fault when people will also start blaming other tools and programming languages for the developers mistakes. However you wont see ars calling out C++ because someone somewhere wrote a buggy program or a program which gives incorrect results.

Excel is not a programming language

 

[rduncan10]

I think Shilmilsh's comments that these are easy mistakes to make in Excel are correct, but I think these would have been easy mistakes to make 30 years ago in the first versions of Excel or Lotus 1-2-3.

It is shocking how little innovation there has been in the spreadsheet application (or any Office application for that matter) over the last 30 years. They've added more rows and columns, made the spreadsheets look prettier, and made a lot of changes (good or bad) to the interface, but at its core, Excel is the same application it was 30 years ago, with all the limitations and benefits it always had.

 

[daarong]

About the insertion error skewing the austerity analysis... really nothing to do with Excel per se. Just like writing software - you can write it so it's rigid, narrow, and the slightest changes break it... or you can write it to be robust.

 

[Jokotai]

ROFL! Work computer background image: changed!

 

[steelghost]

The problem with Excel is that it is easy to get started "writing" an "application". It's the hardboard, pop rivets and duct tape of the office world. If trying to get a proper application (ie supported, with actual infrastructure costs, etc etc) is too hard, people will fill the need with whatever they can cobble together themselves. It may not work well, but it will be better than nothing. Then of course people come to depend on it, forget how it works, and before you know it, there is a cottage industry devoted to maintaining it.

It's a bit like this Swiss army knife, it can theoretically do almost anything, but in the vast majority of cases you'd be better off with something designed for the task at hand.