Timeline for Why do many talented scientists write horrible software?
Current License: CC BY-SA 4.0
54 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Aug 26, 2023 at 23:38 | comment | added | Crimson | It's kind of like doctor's handwriting. | |
| Jun 20, 2020 at 3:14 | comment | added | bob | I wonder if part of the reason might not come from personality differences that lend themselves to very different types of work? Research is highly uncertain, very fuzzy, discovery. Programming is highly discrete, predictable, structured. Generalizations, but still could it be part of the reason? | |
| Nov 21, 2019 at 15:52 | comment | added | SeF | @PeterJansson Playing an instrument has nothing to do with writing code in science. Pillar of science is reproducibility. If the code sucks, the results can not be reproduced, and the scientific method is flawed. | |
| Jul 23, 2018 at 15:18 | comment | added | koverman47 | Even though this is an old post, it is still highly relevant to add that software engineering is not the same thing as computer science. | |
| Jul 22, 2018 at 22:40 | answer | added | gnasher729 | timeline score: 0 | |
| Jul 22, 2018 at 19:59 | answer | added | Fraïssé | timeline score: 4 | |
| Jul 22, 2018 at 17:02 | answer | added | xuq01 | timeline score: 2 | |
| Jul 22, 2018 at 11:01 | review | Close votes | |||
| Jul 23, 2018 at 21:05 | |||||
| S Jul 22, 2018 at 8:57 | history | suggested | Rodrigo de Azevedo | CC BY-SA 4.0 |
Minor improvements
|
| Jul 22, 2018 at 8:40 | review | Suggested edits | |||
| S Jul 22, 2018 at 8:57 | |||||
| Jan 19, 2016 at 0:26 | comment | added | Captain Emacs | I've written code that followed (and, at that time, invented some of what later became) best practice in terms of Software Engineering. And I've written pure research code. When you explore an area, you do not plan out the code, because you do not know what direction will deliver. Planning out code for an unexplored direction is like premature optimisation. You will spend 95% of your coding on something that is used only in 5% of your study. It's wasteful. Only after you explored the area and know what you really consistently need, you can create beautifully clean code. | |
| Jan 25, 2015 at 2:23 | comment | added | Thorbjørn Ravn Andersen | Research code do not need to be production ready, and frequently do not need to maintainable either. That alone gets rid of 90% of the "best practices" for developers. | |
| Jan 23, 2015 at 19:28 | comment | added | reo katoa | It seems that most answerers here are taking it for granted that if a researcher produces a "bad" piece of software, that means they are bad at writing software. You wouldn't look at a lab notebook and conclude the scientist was an incoherent writer with no eye for layout. Writing good software takes time and effort. In many cases, it is simply not a priority, and in many cases, it should not be. | |
| Aug 1, 2014 at 5:24 | answer | added | Greg | timeline score: 7 | |
| Aug 1, 2014 at 4:56 | comment | added | Greg | Why programmers suck UX/UI? Same problem: people think they are smart, so don't spend 5-10 hours for years a day to polish a skill that are totally not there main job, that not obvious they suck at and hard to actually learn. Why would a non-programmer ever hear about version control or different kinds of testings? You know how much testing sucks for sci software? | |
| Aug 1, 2014 at 2:08 | answer | added | Tom Au | timeline score: 1 | |
| Jun 16, 2014 at 14:05 | comment | added | Peter Jansson | Many talented scientists are also poor at playing an instrument, but they still do it. Talent in one area does not automatically translate into talent in all (although some seem to think so). | |
| Jun 15, 2014 at 7:14 | answer | added | user168715 | timeline score: 18 | |
| Jun 13, 2014 at 17:31 | comment | added | Quora Feans | I'd guess that they didn't wrote the code themselves. So, the question should be 'why do underpaid post-docs write so poor code for their professors? Is it just because they don't get recognition? | |
| Apr 29, 2014 at 5:19 | comment | added | dmckee --- ex-moderator kitten | I should note that there are fields---like experimental particle physics---where a non-trivial fraction of students, postdocs, and professors care about programming and programming process. That's not to say that we work up to industry standards, but we make the attempt. Version control is ubiquitous (and moving to distributed systems); big tracking databases and automated builds are widespread; controlled release processes are to be found here and there. Alas code review is still pretty rare. | |
| Apr 28, 2014 at 13:58 | comment | added | Frames Catherine White | I suspect this is closely tied to "Why do my CAD programs have terrible User Interface?" Because they are made by Engineers not Software Developers. (I know I've used $10,000 pieces of software that had no multiple undo, truly unorthodox mouse interaction and was made in a terrifying mix of swing, AWT and C++. | |
| Apr 28, 2014 at 11:31 | history | protected | aeismail | ||
| Apr 28, 2014 at 7:38 | answer | added | algorithmic_fungus | timeline score: 15 | |
| Apr 26, 2014 at 18:08 | comment | added | David Richerby | Why are some people who are good at X bad at Y? Quite simply because X and Y are different things. | |
| Mar 14, 2014 at 20:14 | comment | added | Jeff Dun | Some of the points listed in OP sound very familiar to me. I particularly encounter poor matlab code on regular basis. | |
| Mar 12, 2014 at 13:34 | comment | added | Penguin_Knight | My flippant answer: More concerning, why are many talented scientists a bad kisser? Since those people are very good in doing their research - and eventually obtain remarkable results - it seems they are clever enough to put some mind into kissing. | |
| Mar 12, 2014 at 7:38 | comment | added | Mark Meckes | My flippant answer: Why do many talented programmers write horrible documentation for end users? | |
| Mar 12, 2014 at 6:01 | answer | added | TomRoche | timeline score: 5 | |
| Mar 11, 2014 at 4:15 | comment | added | David Ketcheson | It's worth mentioning that there are very many scientific codes that are quite well-written, employ good practices, etc. I will grant that they are not the majority. | |
| Mar 10, 2014 at 19:23 | comment | added | Fomite | My flippant answer: Can you name one person who got tenure because of their well-written code, or their careful git commits? | |
| Mar 10, 2014 at 19:22 | comment | added | Fomite | @DavidKetcheson I've found that's not particularly true (the GUIs thing). It's been a push at several places I've been to put GUI or web-based front-ends on code to allow for rapid prototyping and the like. | |
| Mar 8, 2014 at 20:18 | comment | added | Iguananaut | Just thought I should add that, as I've found over the years, academics in Computer Science also by and large write terrible software--they do not, after all, work as programmers much of the time. This is of course not a rule--just an observation I think many have made. | |
| Mar 8, 2014 at 17:35 | comment | added | David Ketcheson | There's some relevant discussion in this thought piece: Making reproducible computational research a reasonable choice for young faculty on tenure track. | |
| Mar 8, 2014 at 17:30 | comment | added | David Ketcheson | Scientific codes with GUIs? That is very unusual. | |
| Mar 8, 2014 at 16:37 | answer | added | Stian Håklev | timeline score: 7 | |
| Mar 8, 2014 at 10:09 | answer | added | posdef | timeline score: 30 | |
| Mar 8, 2014 at 8:06 | comment | added | Flyto | A number of answers have mentioned the Software Carpentry project, which aims to teach software development good practice to scientists. You may be interested in a couple of papers that they have published: f1000research.com/articles/3-62/v1 and plosbiology.org/article/… | |
| Mar 7, 2014 at 20:53 | history | edited | badroit | CC BY-SA 3.0 |
fix typo in title
|
| Mar 7, 2014 at 17:49 | answer | added | Piotr Migdal | timeline score: 23 | |
| Mar 7, 2014 at 17:29 | answer | added | TwoThe | timeline score: 11 | |
| Mar 7, 2014 at 13:03 | review | Close votes | |||
| Mar 7, 2014 at 18:20 | |||||
| Mar 7, 2014 at 1:38 | history | tweeted | twitter.com/#!/StackAcademia/status/441749664950661120 | ||
| Mar 7, 2014 at 0:28 | answer | added | cbeleites | timeline score: 87 | |
| Mar 6, 2014 at 20:57 | history | reopened |
Piotr Migdal Faheem Mitha John Doucette Nobody David Z |
||
| Mar 5, 2014 at 21:11 | history | edited | Thanatos | CC BY-SA 3.0 |
edited body
|
| Mar 5, 2014 at 21:03 | history | edited | Thanatos | CC BY-SA 3.0 |
added 13 characters in body; edited title
|
| Mar 5, 2014 at 20:11 | review | Reopen votes | |||
| Mar 6, 2014 at 0:09 | |||||
| Mar 5, 2014 at 20:03 | history | edited | Piotr Migdal | CC BY-SA 3.0 |
serious edit/rewrite
|
| Mar 5, 2014 at 19:54 | history | edited | Piotr Migdal | CC BY-SA 3.0 |
serious edit/rewrite
|
| Mar 5, 2014 at 19:18 | history | closed |
Matthew G. Peter Jansson 410 gone gerrit Lev Reyzin |
Needs more focus | |
| Mar 5, 2014 at 19:01 | answer | added | Marc Claesen | timeline score: 144 | |
| Mar 5, 2014 at 18:41 | review | Close votes | |||
| S Mar 5, 2014 at 19:21 | |||||
| Mar 5, 2014 at 18:31 | review | First posts | |||
| S Mar 5, 2014 at 19:21 | |||||
| Mar 5, 2014 at 18:15 | history | asked | Thanatos | CC BY-SA 3.0 |