Nuances of comparable

Question

English is not my native language and I am unsure of the nuances of the word "comparable".

In the sentence:

It is not expected that the power-effectiveness will equal the effectiveness of dedicated hardware, but it is hoped that it can be comparable.

How similar do the two values that are compared need to be for them to be "comparable".

"within the same order of magnitude" has been suggested, but my values are within 250% of the original value, and I feel that "order of magnitude" is too pessimistic. Any suggestions for a different phrase/word I can use?

Answer 1

You are not going to get a quantitative definition of comparable.

Comparable has so many nuances that it can be anything you want it to be, and anything your reader reads into it. In your sentence

It is not expected that the power-effectiveness will equal the effectiveness of dedicated hardware, but it is hoped that it can be comparable

Replace "....it is hoped that it can be comparable" with an unambiguous, quantitative statement.

In one of your comments, you say: "The power consumption of my design is within 2.5 times the power consumption of the dedicated module.... "within" because it varies with frequency, 2.5x is the worst case."

Thus, I suggest: "The power consumption of my design varies with frequency, and, at worst, is 2.5 times the power consumption of dedicated hardware."

You have summarized concisely and without ambiguity. Forget "comparable" and "order of magnitude", which you are going to have to explain later anyway.

Edit in response to Hida's comments about expectations

I suggest that you present your expectations in terms of "acceptable", rather than "comparable". You decided to build this piece of equipment instead of buying it because...why? It was cheaper? Possibly by building it yourself you improved some key aspect of its performance at the expense of other aspects you didn't care about? If so, try something like: "The equipment that I build was much less expensive than off-the-shelf equipment would have been; it [did whatever it did] as well as (better than? nearly as good as ?) the commercial product would have, although with an increase in power consumption, which could be held to an acceptable level."

You are going to explain all this quantitatively later. Acceptable conveys the message that you had quantitative criteria, which you met. Comparable doesn't have that connotation, to me, anyway. You define what is acceptable; anyone can define what is comparable.

Answer 2

There seem to be two crucial points to consider. 1. Even if 2.5 is the max increase of consumption, the more informative value is the expected new value (understandably, it may be unknown). 2. Who the audience is, and consequently, how precise (or vague) you need to be.

I believe "within the same order of magnitude" is actually one of the more flattering ways of putting things, if it can be expected that the new values will be mostly about double of the original ones. The reason is simple -- if consumption increases twice, then only half of all the possible values will be within the same order of magnitude (in decimal, of course). If the increase is more than double, then a statement about the same order of magnitude is not even true.

Comparable or proportional are vague, placating, and (in their strict sense) vacuously true terms that may perhaps be considered if there is no risk of alienating the audience by being vague, or if the power consumption is a side issue, etc.

Added: The above was written understanding order of magnitude as described here (i.e., as equivalence, where each class is given by the exponent of 10 in the scientific notation of that number). Apparently though, that is not the prevailing meaning of that notion; many interpret it as a relative notion, cf. John Baez quoted in the Wikipedia entry.

Answer 3

I would posit that that this is an answer dependent on your target audience.

There are two kinds of target audiences (generally speaking) Professional, and Layperson. For my purposes, Professionals are people with detailed knowledge of the subject under discussion, and Laypersons are not. They may be Professionals in other spaces but they lack depth of knowledge and breadth experience on the subject relevant to the writing. For both, comparable means that there is a notable difference relevant to their knowledge. For example, with your module:

Would a Professional [above definition] find cause for pause when deciding to purchase your module over the dedicated module due to it's power consumption? Would they have to modify other parts of a project including your module due to these increased power requirements? If so, it is not comparable.

Would a Layperson even notice? If not, it is comparable.

Answer 4

You're asking for a quantitative expression of a qualitative term. It's like asking how many notes two songs may differ by before they stop feeling similar. You'll need some other reference to make the comparison.

In the case you present, whether two values are comparable should be determined with reference to practical usage. For example, if the devices are solar powered and each device requires only a single solar panel, you might say their power requirements are comparable. If one needed two solar panels instead, you might say their power requirements weren't comparable.

Your question of

how similar ... two values that are compared need to be for them to be "comparable"

can be answered by as much as the system can bear before (significant) additional resources are needed.

The final question in your post was

Any suggestions for a different phrase/word I can use?

If you are asking for a phrase in place of comparable, try within engineering tolerances. This means that you set limits (e.g. one solar panel), and then check whether those limits are breached. Those limits are called tolerances. If the limits aren't breached, the system is said to be operating within tolerances.

It is not expected that the power-effectiveness will equal the effectiveness of dedicated hardware, but it is hoped that it can be within engineering tolerances.

Answer 5

You started with

It is not expected that the power-effectiveness will equal the effectiveness of dedicated hardware, but it is hoped that it can be comparable.

As mentioned by others, power-effectiveness isn't correct, nor is it commonly used. We can also rearrange this to make it less awkward.

It is expected that the power-efficiency will not equal the efficiency of dedicated hardware, but it is comparable.

Comparable here doesn't have a context. For engineers you would simply give the numbers, so I'm assuming this is marketing speak. In that case, we can spin the phrasing to emphasize the positive.

The [unit] only requires 2.5x more power than dedicated hardware.

From this, someone understands that dedicated hardware is a gold standard, that 2.5x is something to brag about. Maybe we want to take one step further, and add complexity to the statement to show that 2.5x is a worst case scenario. We can also add some guidance on the standard. This is optional, and starts sounding less technical and more like marketing--but may be more useful in your case.

The [unit] only requires 2.5x more peak power than highly efficient dedicated hardware.

Answer 6

Comparable literally means, able to be compared.

In current times, the meaning has shifted slightly and in general use it now means something closer to of near-equal value.

In the quoted sentence, it is more of a negative expressed as a positive. In other words, the writer is saying, "If not the same, at least they will not be very much different".

However, this can also be a matter of perspective. A cow is reasonably comparable to a bull. However, to an alien, a cow may also be comparable to a horse, or even to a camel.

If your results are 250% greater than what you are comparing against, then they are least comparable when contrasted with results 20,000% greater. Otherwise, perhaps you can say that your results are somewhat higher than anticipated, or that they are larger by a certain percentage, or within acceptable range of deviation.