1

In Proper verb to denote 'consistent-ize', the top answer by vote says "standardize" should be used, but I have a few doubts.

Firstly, English is not my mother tongue, and I use it predominantly in programming.

So my understanding for the word "standardize" is that it mostly mean to formalize a set of rules, practices, and interpretations for the purpose of improving quality and interoperability of products.

My intended usage of the verb form of 'consistent' is as follow:

Make the content of a dataset conformant to the rules set in a schema.

So, in a technical sense, "standardize" isn't the correct word for it. And "normalize" is close to what I'm looking for, since my understanding for the word is mostly in the mathematical-statistical field, but there's something "consistent" is more appropriate than "normal", for example:

2+2=5

I can make the formula consistent by changing a number, or the comparison sign, but it doesn't make any sense to "normalize" it. It fits my sense of "consistent-ize" where the tuple (2,2,5) is the data set, and a+b=c is the schema.

In a comment I see "regularize" being mentioned and up-voted, and it fits the best.

So my question is: can I say "regularize an inconsistent statement/equation/predicate" using the technical sense of the words? Is there no better word to mean make something consistent according to the rules?

3
  • Try compliant. Good Luck.
    – Kris
    Feb 25 '19 at 11:10
  • @Kris Wiktionary and Google Translate returns it as a adjective. Are you sure it can be used as a verb?
    – DannyNiu
    Feb 25 '19 at 11:46
  • DannyNiu, you already have make so you only need a noun, not a verb.
    – Kris
    Feb 27 '19 at 9:24
1

I like harmonize (user22542's answer) as a generic verb that means to make two things conform with one another. But you seem to be asking for a more specific term relating to data.

Your phrase "Make the content of a dataset conformant to the rules set in a schema" implies that the rules are being retrofitted onto the dataset, that is, these rules were not in effect as a "screen" to prevent rules violations as the data were being inserted. Data-validation usually refers to the prevention of rules violations, not to a clean-up after the rules have already been violated.

Data that violates certain kinds of rules can be sanitized to make it rules-conformant.

4
  • In programming, there are syntax errors that causes compilation failures, and logical bugs that causes successfully compiled program to misbehave. In the same way web programming (my field) have sanity having to do with security and consistency having to do with application business logic. And so I should be prefixing "sanitize" with "security-" or "logical-" if I want to refer to the two distinct subtopic, am I correct?
    – DannyNiu
    Feb 25 '19 at 14:10
  • @DanyNiu: Are you asking about some conformity check that makes sure your procedural code properly implements and properly enforces a set of declarative business rules?
    – TRomano
    Feb 25 '19 at 14:23
  • Yes. I was actually hoping for separate words for security and business logic. "Sanitization" is too familiar a word in our security department.
    – DannyNiu
    Feb 25 '19 at 14:31
  • 1
    In that case, you're no longer talking about "datasets" and I think harmonize is probably the best verb: We have to harmonize the middle tier logic with the business rules.
    – TRomano
    Feb 25 '19 at 16:13
1

The word "harmonize" seems to agree with your description and intended usages. It was offered as a brief answer buried lower in the 'consistent-ize' link that you referenced, but it was not very conspicuous. One common meaning given is: To bring or come into agreement or harmony.

I would like to "make the content of a dataset "harmonize" to (with) the rules set in a schema.

or

How many ways can one "harmonize" the data set (2,2,5) with the a+b=c schema?

https://www.thefreedictionary.com/harmonize

1
  • Although the reference didn't mention anything to do with "consistency", I believe the word had been used in such sense in quite a few academic and technological documents.
    – DannyNiu
    Feb 25 '19 at 13:07
0

It has been difficult to find a transitive verb wich expresses what you are doing, there are many such as comply which are intransitive but they need the cooperation of the data. For example "The data comply with the schema". However the Oxford Online Dictionary defines the verb constrain as

VERB [WITH OBJECT]

1Compel or force (someone) to follow a particular course of action.

1.1 Severely restrict the scope, extent, or activity of.

You could describe the action of making the data conform to the schema as constraining the data. This would cover the process of checking for conformity and editing the data until it actually conformed. Obviously there are dangers in using code to amend data in this way, but the word does describe what you are doing

2
  • Would you provide advice on potentially using non-dictionary words such as "validify" or "validatize"? I see them a potential bonus points in an answer.
    – DannyNiu
    Feb 25 '19 at 12:00
  • 1
    @DannyNiu This site deals with English language and usage. It is, of course, possible to invent words, they are called neologisms but they aren't considered to be part of the language until they have been used by a reasonable subset of the population. For example someone was the first to talk about taking a selfie or being hangry and these words are now in at least one respected dictionary, but only because they are now widespread. You could suggest a neologism on this forum but it would not be a definitive answer.
    – BoldBen
    Feb 25 '19 at 13:09
-2

If you're looking for A purely technical term and don't care if "muggles" can understand, then

lint

can be what you look for.

In Single Unix Specification, it's mentioned lint is a program that checks for potential errors in a C language program source code. The specification did not specify the behavior of the program as the standard developers find it out of the scope.

Nowadays, lint refers non-specifically to a class of programs that checks for potential errors in textual or binary coded information set, such as es-lint that works with JavaScripts.

6
  • I don't think lint would work, Danny, my understanding of lint is that it is, essentially, a syntax checker for code. The OP is asking about ensuring that data conforms to the specification of schema, that's a very different thing.
    – BoldBen
    Feb 25 '19 at 11:09
  • @BoldBen Example of lint that isn't a programming language syntx checker: github.com/tanepiper/npm-lint
    – DannyNiu
    Feb 25 '19 at 11:15
  • @BoldBen Example of lint that checks AI training data: github.com/brain-research/data-linter and it checks more than schematic conformance.
    – DannyNiu
    Feb 25 '19 at 11:25
  • 1
    I see that it's more than a simple syntax checker, Danny, but from my understanding of the description it's still verifying that the operation of the json file is producing consistent, secure and compliant results, not checking and correcting the structure of existing data. I think that the OP is looking for a word for a process closer to checking the structure of the data in an XML file.
    – BoldBen
    Feb 25 '19 at 11:26
  • 1
    Spotting errors is not the same as fixing them, and you asked about making the data conform to a set a rules, not merely about identifying where they violate the rules.
    – TRomano
    Feb 25 '19 at 13:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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