Terminology: Difference between decision variables, features and attributes?

Could there be a difference between the words “feature”, “attribute”, and “decision variable” when used in the same paper? The one I am specifically thinking about is about an optimization method for clustering, but I am also wondering if there generally are any scenarios for which it could be.

I can’t manage to google up an answer that either confirms or denies that these are the same thing, and I have no formal training in data science.

Terminology for reduced brute-force attack?

Many security algorithms today have such a large key length, that there’s just no use in trying to brute-force a key. For example to find one AES-256 key you would have to try 2^128 keys on average.

My question is, if there’s a special name for a brute-force attack where you would somehow know that a key “lies” in a specific range, which would reduce the amount of calculations so drastically that a brute-force attack suddenly becomes feasible.

An example about simple brute-force RSA-factorization might make it more clear:

We have a 2048-bit RSA key. Way too large to just try one number at a time to brute-force it.

But we assume now that I have a special algorithm that doesn’t directly return one of the factors, but can tell us that one of the factors lies in a certain range k (plus / minus 1,000,000,000). Then we would only have to try the possibilities from k – 1,000,000,000 up to k + 1,000,000,000 and this is probably quite possible to brute-force.

No matter if such algorithms exist or not, is there a special name for this specific attack (reducing the possible key range to make a brute-force attack feasible)?

Terminology: Blue-Green deployment

Let’s say your organization uses a blue-green deployment model. Simple question:

  • After Blue has gone live, do you rename it to Green, and the server(s) formerly known as Green now becomes Blue?

Or to put it another way:

  • Are “Blue” and “Green” the names of the physical environments? Or are they logical names, i.e. synonyms for “current primary” and “current secondary”?

Correct terminology for if-blocks

I am using lex/yacc to parse a simple language which allows ‘if blocks’:

IF something    something   [ELSE IF something][ELSE] END IF 

and also other nested logic under a ‘DEFINE’ statement:

DEFINE something   something   something else END DEFINE 

I am in need of a common term to refer to these blocks (as the name of a container class). I have thought of Flow Control but it does not seem to correctly capture the intent…?

What is known about tabaxi language and terminology, particularly the word “tabaxi” and the name of the Cat Lord?

According to this Dragon+ article, the term tabaxi originates with the race of cat people (at least in the Forgotten Realms setting as of D&D 5th edition):

Tabaxi with a capital T is the ancestral name of one of a number of human tribes that traveled east to Faerûn from across the ocean…

…although having both tabaxi and Tabaxi in the canon of the Forgotten Realms is confusing, it makes sense for tabaxi to originate from Maztica. This provides the in-world explanation for the similarity of their names: the human Tabaxi named their tribe after the mysterious cat people of Maztica and Katashaka. Also, establishing them as native to that distant land explains why tabaxi are rare across Faerûn in both the past and present.

But nothing further is explained about the name beyond that. I haven’t found any other information about the name of the race, but I am curious what other information there is, e.g. whether the word “tabaxi” ultimately originates from the cat people, and whether it has any meaning other than “this specific race of cat people”.

Likewise, it seems as if the tabaxi should have their own name for their figure of worship, aside from the plain phrase “Cat Lord” in Common. This answer explains that the first Cat Lord in D&D was named Meerclar, but that pre-dates the creation of the tabaxi and the Fiend Folio doesn’t mention any figure of worship. (and the 2nd edition books seem to indicate that tabaxi religion is largely animistic.) Likewise, it’s mentioned that the corresponding figure in Greyhawk is named Rexfelis, but there don’t seem to be cat people in Greyhawk and the Latinate name seems like a poor fit for the tabaxi.

One might presume that “tabaxi” means something like “cat people”, and that the Cat Lord could have some related name in the tabaxi’s language, but is there any established information about these subjects in any official D&D publications?

Terminology for explaining a limited range of values encoded into bits [on hold]

It’s been incredibly difficult to do searches and write essays on subjects like these without knowing the proper terminology.

Here’s an example, to clarify what is being discussed here:

The four nucleotides – the base units or “letters” of DNA – are Adenine, Cytosine, Guanine, and Thymine. 4 different possibilities means you can store each nucleotide in 2 bits.

  • A – 00
  • C – 01
  • G – 10
  • T – 11

It took forever to find again, but ASN.1 is the kind of “encoding” I’m talking about. Layman’s Guide to a Subset of ASN.1, BER, and DER.

Elusive terms:

  • The general name for this.
    • A search for “byte encoding” returns only UTF-8 and ASCII stuff.
    • The term “serialization” includes string formats like JSON and XML, which are not what is being discussed.
    • The Wikipedia page for serialization even has to refer to this as “the more compact, byte-stream-based encoding”.
  • Term to generally refer to one of these non-human-readable streams of data, in a specific format.
    • “File format” describes what I’m talking about, but it’s difficult to convey the difference between an FTP JSON file and a “custom-serialized” BLOB on an SQL server.
    • “Codec” describes it well, but it is referring almost exclusively to video and audio.
  • Term to refer to a domain of possible values to be encoded into bits.
    • For the DNA example above, this would refer to the 4 unique letters; the members of an enum.
    • For a numeric value, the range would determine how many bits are required to represent the number.
    • For a string value, the range of characters available would determine how many bits are required per letter. Variable vs fixed-length string would also be a part of it (covered in the ASN.1 guide).
  • This field as a whole.
  • Any others that potentially would be helpful in the future.

And if there aren’t specific terms available, then provide alternative ways to convey these concepts.

Is there more apt terminology to research: order management systems comprised of a hierarchy of entities?

I’ve been grappling with a project at work that seemed straightforward at first but then began to spring up increasingly complex corner cases. Over time I noticed many similarities between this project and one at my previous job, which was basically an order management system (OMS). With how common OMSes are, I feel someone must have exhaustively solved every kind of problem that OMSes can have, from an abstracted system design perspective. But I’m having trouble finding any such reference, using keywords like:

system design architecture order management system oms version mismatch 

The specific subproblem I’m trying to gain insight on is this:

  • You have orders that you send to an external Source of Truth (SoT), be it buy and sell orders on a stock exchange, or line items on an ad server.
  • You provide an interface for your customers to create and manage their orders that get routed to the SoT.
  • Orders have many, many states, affectable by either the customer or the SoT, and they are also versioned but not necessarily audited (i.e. you can’t retrieve the order as it looked N versions ago).
  • Message crossovers are possible. (This is the heart of the matter.) Either there are periods of rapid activity (e.g. opening bell on the stock exchange) that literally cross over with traders’ actions that were made with stale information, or actions take time to formulate or review (e.g. a UI where account managers review the details of an order to approve it) but during that time the order in question may have changed (e.g. edited by the customer, etc.)

How do you design a system that makes sure nothing unexpected happens, at least not without sufficiently notifying the user (and if applicable, reconciling with the SoT). Is there a name or reference for this specific genre of problems? To give very specific examples:

  • There should be an error if the customer attempts to edit the non-latest version of an order.
  • But if the customer’s edit and the latest version are identical, it can be useful to silently succeed (e.g. user can apply an action from different views in the UI; opens a browser tab and commits one action, then in another tab commits another action that implies the first action).

Actually, up to this point, I can handle / come up with a system design myself. But there’s an added dimension (literally) that, in my experience, makes both systems extremely complex and error-prone:

  • Orders are actually a hierarchy of entities, each with a unique set of states and rules, though all versioned. For the stock exchange, the OMS may have had a concept of “parent orders” that housed “child orders.” Worse, the ads OMS has a five-layer hierarchy. What kind of problems do hierarchies introduce? Well, if editing a parent order cascades edits to its child orders, but one the child edits fail at the SoT, what is the state of the parent? If the serving state of an ad depends on all 5 layers being “approved,” and one layer has been edited, … you get the idea.

Is there not a named discipline or textbook that covers: Designing an Order Management System of a Hierarchy of Versioned Entities. That’s what I’m hoping to find out how to research. Originally, I thought maybe I should try to write that reference. But after a few brainstorming sessions, I think I’m in way over my head, and it just seems there’s no way someone didn’t already think all variations of this problem through.

Question regarding terminology used to describe Benes networks

I’m studying a routing problem on a cyclic butterfly network. The paper I am reading states that two properties of the cyclic butterfly network make it efficient for the routing algorithm it describes. I’m new to this subject, and I’m struggling to understand the first property, which is:

“(1) It embeds a Benes network – meaning that if we traverse the columns and back any permutation of the row labels can take place without collisions”

It adds that this is trivially true for a square grid. In this context do collisions mean collision of packets of information? Why is this trivially true for a square grid, but not the cyclic butterfly network?

Kind regards