Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

One in five genetics papers contains errors thanks to Excel (2016)

science.org

Post reply on HN
Print view View original

21–30 of 267 posts

Re: One in five genetics papers contains errors thanks to Excel (2016)

#21

Post by TeMPOraL »

happytoexplain wrote: post #9
b112 wrote: post #3

It's not really Excel, is it? But users not reviewing what's happening, not paying attention.

This attitude is at the root of why software is so much worse than it could be. "It's the users' fault" as a real root cause is borderline non-existent, because it can always be obviated in the software, even though the kind of talent that can do that for complex systems is exceedingly rare.

I disagree. The dominating attitude today is that users are idiots and too dumb to use anything they can't master in 30 seconds from first exposure, which has a nice side effect of cutting out a lot of functionality from the scope. The result is, for example, 20 startups going after any given idea, each spending years refining a slightly different take on a useless set of simple features.

(Actually, many of such products can be made better by replacing them with an Excel sheet, which is a big part of the reason why people who actually need to get shit done end up using Excel.)

Re: One in five genetics papers contains errors thanks to Excel (2016)

#22

Post by nitwit005 »

happytoexplain wrote: post #9
b112 wrote: post #3

It's not really Excel, is it? But users not reviewing what's happening, not paying attention.

This attitude is at the root of why software is so much worse than it could be. "It's the users' fault" as a real root cause is borderline non-existent, because it can always be obviated in the software, even though the kind of talent that can do that for complex systems is exceedingly rare.

This is users using software in a way the designers never expected, decades after it was designed. Those users then chose to ignore warnings about using it.

Newer programs like Google Sheets have better default behaviors. They have free access to it.

Re: One in five genetics papers contains errors thanks to Excel (2016)

#23

Post by TeMPOraL »

rectang wrote: post #8
HenryBemis wrote: post #6

Let me reword this. "One in five genetics papers contains errors because the authors were careless when using Excel". I dislike Excel for what it does (EUCs, nearly impossible to track changes, etc.) But on the other hand it is an amazing tool.

Oh, come on. The Excel user interface encourages mistakes, as surely as a bicycle with handlebars that steer backwards. https://www.youtube.com/watch?v=MFzDaBzBlL0 It's not all on the users.

At least it's a bicycle.

Remember the whole "computer as bicycle for the mind" thing? That didn't happen, the world went in the opposite direction. Software like Excel are the last surviving remnants of the idea of empowering end users to improve their work and lives.

Re: One in five genetics papers contains errors thanks to Excel (2016)

#24

Post by TeMPOraL »

rectang wrote: post #20
nitwit005 wrote: post #14

Earlier quoted context omitted.

This isn't Microsoft lecturing them. It's entirely valid for the school to tell students to avoid easily avoidable pitfalls.

Of course they can lecture them — it just isn't going to work. The error rates will barely budge. There is one effect: it allows smug gloating about how stupid, lazy, and irresponsible these users are. Blaming the user is the last refuge of the incompetent.

> Blaming the user is the last refuge of the incompetent.

Why then is it the dominating mindset in software design today, and advertised as being the opposite to the mindset that gives you Excel?

Re: One in five genetics papers contains errors thanks to Excel (2016)

#26

Post by chucksmash »

prepend wrote: post #15
monocasa wrote: post #7

Earlier quoted context omitted.

> One in five genetics papers contains errors because the authors were careless when using Excel Or Excel is a remarkably easy tool to mishandle because of generally unexpected transformations it makes 'for you' automatically in an easy to miss way.

You think Excel should make special accommodations because genetics authors misuse it? Excel has been acting this way since before bioinformatics existed. Authors need to use their tools properly.

Excel (and GSheets) are tools where I think a stripped down "trust my input, I know what I'm doing" mode that doesn't try to intuit user intent would be useful.

As I type this comment and my phone miscorrects "intuit" to "Intuit," I think also Google keyboard could benefit from such a mode that only handles spelling mistakes but doesn't replace uncommon words with common brands, etc.

Re: One in five genetics papers contains errors thanks to Excel (2016)

#27

Post by saagarjha »

prepend wrote: post #18
rectang wrote: post #11

Earlier quoted context omitted.

Haranguing your users not to make mistakes is utterly worthless. The only way to move the needle is to design systems that are easier to use correctly.

Excel is not a genetics tool. It has millions if use cases, many more important than genetics. Excel didn’t care about this. It’s like people complaining because sugar gets misused. Or that murderers stab people with knives. The solutions isn’t to “fix” knives.

Excel is a tool for general-purpose data processing. What the genetics people are doing is exactly that.

Re: One in five genetics papers contains errors thanks to Excel (2016)

#28

Post by happytoexplain »

TeMPOraL wrote: post #23
rectang wrote: post #8

Earlier quoted context omitted.

Oh, come on. The Excel user interface encourages mistakes, as surely as a bicycle with handlebars that steer backwards. https://www.youtube.com/watch?v=MFzDaBzBlL0 It's not all on the users.

At least it's a bicycle. Remember the whole "computer as bicycle for the mind" thing? That didn't happen, the world went in the opposite direction. Software like Excel are the last surviving remnants of the idea of empowering end users to improve their work and lives.

> Software like Excel are the last surviving remnants of the idea of empowering end users to improve their work and lives.

I agree. I love Excel. But this attitude only makes sense if we assume one can only fix easy-to-make mistakes by dumbing the software down, which is not true.

Re: One in five genetics papers contains errors thanks to Excel (2016)

#29

Post by Prickle »

Previous popular thread, Aug 2020: https://news.ycombinator.com/item?id=24070385

> Scientists rename human genes to stop MS Excel from misreading them as dates (theverge.com)

Related details:

2023:

https://www.pcmag.com/news/microsoft-finally-fixes-excel-gli...

> Years after introducing Excel's automatic conversion features, Microsoft rolls out an update to prevent it from changing gene symbols to dates.

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9325790/

> Gene Updater: a web tool that autocorrects and updates for Excel misidentified gene names

Re: One in five genetics papers contains errors thanks to Excel (2016)

#30

Post by magicalhippo »

I use Excel a fair bit at work, mostly for looking at data, but also to fairly easily generate SQL statements based on such data, often after some formulas have been applied.

Though I think what I'd really want is some tool which has the same grid-like visualization, filtering and direct entering, but was code-based under the hood and without magic conversions.

So, take Excel, and when you enter a formula in a cell, it actually writes a line of code for you, which you can inspect and edit. Including adding your own functions and such.

When importing delimited text data, you'd have to specify what the data is in each column. It should still save the original text data so you can change your mind, but yeah, no automagic stuff.

Post reply on HN