14

I'm working on a data model which, I think, produces more accurate and valid results than some of previous works by others, including my supervisor, who by the way supports my work. But so far, the relevant community in general hasn't paid enough attention to the need for such more accurate results in this field.

When presenting my work (verbal/poster/paper), I need to emphasise the importance of the problem and the need to using better models, and that if we used less valid models (such as some of the existing ones) we should expect less accurate results.

Is it rude to use the computer science axiom "garbage in, garbage out" (or similar phrases) when referring to others' works in this context?

Edit: I missed important details. In my case, by 'garbage' I specifically mean the low-quality data usually used to solve the problem, which makes other models less valid. Even though, I would never use it as it might be misinterpreted as evaluation of the works per se, rather than the resources used!

But I was surprised and not comfortable to read the analogy in a published comment which criticised another author's results for using unreliable data!

  • 47
    Yes, it would be rude, unprofessional, and unfunny. In short, don't do it. – Dan Romik Oct 25 '15 at 21:14
  • 38
    Yes. This should be obvious. – Corvus Oct 25 '15 at 21:15
  • 5
    To avoid the poor connotations of garbage in garbage out, I often switch to gold in, gold out. Expresses the same idea, the output is only as good as the input and may be improved with better input data. But I'm not sure I'd consider using GIGO in either formulation in a formal paper outside of CS. – user0721090601 Oct 25 '15 at 23:26
  • 21
    How could you possibly think that it might not be rude to call somebody else's work "garbage"? – David Richerby Oct 25 '15 at 23:41
  • 14
    "Garbage in - garbage out" is true of any model or algorithm (e.g. yours too), thus it is not a criticism of a given model or algorithm. So it's not rude, but it's not usually used the way you intend to. – Cape Code Oct 26 '15 at 7:16
44

Yes, it's rude to describe somebody else's work as garbage, and unnecessarily so. You've already said it much better yourself:

if we used less valid models we should expect less accurate results

There's no reason to confuse this simple factual statement with a value judgement. Just because something works poorly doesn't mean it's garbage---it may also simply be earlier and less well engineered. Would you call a Model T "garbage" just because it gives inferior performance compared to a current car?

  • 13
    Your logic is... missing something. The conclusion you want is "don't call stuff garbage; it's bad for your career." You say it's bad because it's rude; fine. But you then talk about value judgments, as though being a (negative) value judgment makes something rude. And then you talk about how old things might just be worse and therefore the value judgment "garbage" is erroneous to begin with, but this also doesn't follow; some researchers' work really is garbage, and it will remain so 80 years from now. – user4512 Oct 26 '15 at 2:52
  • 2
    Debate on value judgements is pointless, so people try to stick to facts to avoid endless debates. There's certainly garbage around, but people won't agree on which, unless it's such garbage that it's not worth mentioning. – Blaisorblade Oct 26 '15 at 10:37
  • @ChrisWhite I have a deeper problem with an insult like "garbage," which is that it's not very informative. All "garbage" tells me is that the person using the word doesn't like something, but it doesn't say why. Is it poorly grounded? Badly motivated? Sloppily executed? Just plain incomprehensible? A word like "garbage" is essentially ad hominem for ideas, and just as poor an argument there as ad hominem normally is for people. – jakebeal Oct 27 '15 at 20:54
33

Not only would this be rude, but also downright wrong. Science proceeds by successive approximation producing better and better models. But when we find a better approximation, we don't start calling all the previous models "garbage", for at least two reasons:

  1. Out of respect, because we know that we wouldn't have found this new model without all the previous ones. And should someone else find a better model than the one devised by us, we wouldn't like to hear our work called garbage.
  2. A new and better model does not necessarily remove all the previous models, because the domain of applicability may not completely overlap.

Therefore, highlight the qualities of your work, but don't be rude about the work of those who came before you.

18

I think the thing here is that you're conflating the invalid input/invalid output of garbage in garbage out with, well, less accurate models being replaced by more accurate models, which is ENTIRELY a different thing. If you put garbage data in to your model, you would get garbage results out of your model. This does not change no matter how sophisticated the model is.

In short, no, from both a CS standpoint - your model would have the same problem as any other model, or indeed any piece of code, taking invalid input -, and, as others have said, a propriety standpoint, in that you really shouldn't be insulting everyone else's work.

E: Perhaps another way to put this. You're confusing the input with the actual calculations and modeling. If someone input that same guy's data in to your model, it would be equally as shaky in the reliability of the calculations. Accuracy of calculations is another matter entirely.

Another note: I am approaching this solely from a programmer's viewpoint, but the points about not insulting other people are all very good reasons to not use GIGO in your presentation.

E2:Okay, I seem to be serial-editing as I think of stuff.

As regards your edit ("missed important details..."):

I don't get what you're saying there.

Are you saying that you're pushing for more accurate input? Or are you saying something else? Your question and your clarification are at odds.

  • Regarding your "E2" section, requests for clarification of the question should be placed in comments on the question, not in answers. You should have enough rep to start posting comments soon (1 more upvote should do it.) You can find some general information about what's expected for answers in How to Answer. Other information about how SE sites work can be found in the help center. – reirab Oct 26 '15 at 4:05
  • 3
    I like this answer! In GIGO, the first G refers to the input that gets processed in the program. In your case, you're refining a model. This will give you improved results. GIGO isn't remotely relevant. – aparente001 Oct 26 '15 at 5:30
8

I'm going to disagree with some of the other opinions here.

GIGO is a recognized term in the industry. Using it appropriately is, well, appropriate, even if people in other industries wouldn't recognize the term as a very common, and appropriately used, term.

Also, in the real business world, people poke fun at competition in advertisements. I've certainly seen it done before. A lot of academia is designed to prepare you for the business world.

You may brush on some sensitivities, and perhaps even offend some people, while other people may chuckle. In short, your level of success may depend on how well you know your audience. What works well for computer programmers might not work as well for an audience that doesn't recognize the jargon and places a lot of value on things like showing respect for others.

However, even if your audience could appreciate a GIGO reference, you must make sure you're using the term right. As a computer programmer myself, I am familiar with the phrase. "Garbage", in this case, doesn't usually mean "low quality data". It means "wrong data", or "random data that could be wrong, and there is a high probability of (at least some data being) wrong". "Low quality" doesn't really fit that technical description of "garbage". Any craftiness points you might deserve would be offset by the inappropriateness of your simple wrongness.

  • 1
    There is another usage of 'Garbage' in computer science, namely Garbage Collection. In that context, garbage just means the memory that is no longer needed. It's neither low quality data nor wrong data. Its only meaning is "no longer needed". – scaaahu Oct 26 '15 at 12:52
  • 4
    The garbage reference in GIGO refers to bad input data, as your answer points out. If you don't mind me adding a bit: GIGO is not an indictment on the system running the data; rather, GIGO essentially means that, no matter how good the original [software] system might be, if you feed bad data into it, you'll inevitably get bad data out of it. The acronym goes back a long way; I can remember my grandfather explaining it to me in the 1970s. So long as it's being used accurately and appropriately, I think it's okay to use. – J.R. Oct 26 '15 at 15:10
  • 1
    @scaaahu The meaning of 'garbage' in the concept of garbage collection is completely different from and irrelevant to the meaning of 'garbage' in GIGO. – reirab Oct 26 '15 at 19:48
  • I think this is pretty close - I would note that I would only use this in an informal setting though. I wouldn't use this phrase in my root cause analysis document, but that's more because I would be more specific about the why the data didn't work, not because I'm afraid of insulting someone's sacred cow. – corsiKa Oct 27 '15 at 19:38
2

You initially began by talking about models, but in your update it seems you're really talking about collected data. I'm still a bit unclear.

But it seems that you're grappling with uncertainty in different data sets, so I propose that instead of using the computer science term GIGO, you use a signal processing term: noise. All data is noisy, your new techniques give you a less noisy dataset and therefore allow more accurate predictions.

0

It actually depends on if they understand that their work wasn't good enough or look for a better pharse like 'what you put in is what you get out'

0

The golden rule "do unto others as you would have them do unto you" (and the various equivalent statements that appear in a variety of ethical codes) applies just as much to academia as elsewhere. Ask yourself, would I consider it rude if someone (1) referred to my work as "garbage" in the context of GIGO? If the answer is "yes, probably", that suggests it would also be rude for you to use it to refer to the work of others.

I suppose technically that still leaves the question as to whether it is acceptable to be rude to others in an academic paper! ;o)

(1) They say "familiarity breeds contempt", and nobody is more familiar with my research than I am, but there is a difference between being contemptuous of my own work and having somebody else do it for me!

-2

Yes, it is conditionlessly rude. Other answers explain very well, how bad is if you communicate negatively in a hierarchy. I extend this from the other site:

Although it is rude, it can be justifiable. Sometimes people do garbage, it can have many reasons, many of them is tolerable, many of them isn't.

And, naming the things as they are, makes the situation much clear if the reality is masked by complex polite formulas.

It depends on the balance. But, don't forget: people around you will weigh the real hardness of your negative words probably more serious, as you intended it!

Not the answer you're looking for? Browse other questions tagged or ask your own question.