General > General Technical Chat
Admit your Brain lock
coppice:
--- Quote from: Zero999 on April 04, 2024, 03:30:52 pm ---The Latin alphabet is used over much of the globe, for the purposes of information exchange. Until fairly recently, virtually all text was ASCII. A good number of those Unicode symbols look virtually identical, which is also a security vulnerability: look up homoglyph attack. There's no need for that crap in the field of mathematics. It's pure obfuscation. :palm:
--- End quote ---
Unicode is certainly a security nightmare. Between all the bugs in the character set (e.g. Chinese characters split into 2 separate ones that aren't actually different), the ability to express the same string in multiple ways, and other complexities, its nearly impossible to do a simple comparison between two pieces of text in Unicode. IBM produced a massive normalisation library in the late 90s, to try to make strings comparable, but its far from a complete answer, and isn't that widely used anyway.
coppice:
--- Quote from: tggzzz on April 04, 2024, 02:44:12 pm ---In the UK I was impressed about how teaching arithmetic to 7-11 year olds had improved.
In my day it was "follow the algorithm" - which was admittedly useful when I've had to implement floating point arithmetic! Nowadays they are taught that you can get the right answer by several different "successive approximations", e.g. 99y is easily calculated as 100y-y. That encourages a much better "feel" for the "shape" of numbers and arithmetic.
--- End quote ---
The only things we were taught in maths as "follow the algorithm" were the basic arithmetic operations between 5 and about 9 years old. Everything else was derived, explained, and often alternatives were presented. Now the ONLY things taught in a way to develop understanding are the basic arithmetic operations. Everything beyond that is spoon fed. I'm not sure that "follow the algorithm" is a bad thing in primary schools. Until you've started to identify which kids are capable of anything more than following rote procedures, flexibility might be a problem.
Zero999:
--- Quote from: CatalinaWOW on April 04, 2024, 05:01:12 pm ---Absolutely not intentional obfuscation. There are more than 26 unique concepts/entities in math and physics. The people who started the use of these "special" characters did it to provide shorthand and avoid confusion with other entities. The usage was productive and therefore adopted by others, eventually becoming widespread, used by people whose written language did not involve those characters.
In today's world of ubiquitous data processors there is an argument that multi-character identifiers could be used, and commonly is for some symbols like pi. But this all started when you had to write your ideas on paper with a pencil or pen and a terse notation is a benefit to avoid getting bogged down in the mechanics of writing equations. Having a different notation for notebooks/scratch pads and publication made (and makes) no sense so it stuck.
I am unconvinced that long variable names would be an improvement.
Area_of_Circle Is_Identically_Equal_To Distance_Around_The_Rim Times The_Square_Of(Distance_From_Rim_To_Center) Times The_Ratio_Of_Circumference_To_Diameter
This expression still requires prior knowledge of terms like Equal, Identially, Rim, Square, Center, Ratio, Circumference and Diameter and is horribly unwieldy. Imagine the fun for a truly complex expression.
Finally, as a native speaker of English, I will point out the 26 characters is not enough even to encode the languages common in Europe. Even leaving the Greeks and Slavs out there are enough umlauts and tildes and diacritical marks to exhaust that limited set. Unicode may go too far, but was done in a time when it had been found that even the expanded 255 characters in a byte weren't really enough and the obvious idea is to dedicate two bytes.
--- End quote ---
I'm not in favour of long variable names or even the complete elimination of non-Latin glyphs, but a happy medium. There's no reason to use lower case v and upper case V and similar looking glyphs ϵ and E in the same formula. Far from making it easier to write down, it introduces more room for error and confusion.
IanB:
--- Quote from: Zero999 on April 04, 2024, 06:31:23 pm ---I'm not in favour of long variable names or even the complete elimination of non-Latin glyphs, but a happy medium. There's no reason to use lower case v and upper case V and similar looking glyphs ϵ and E in the same formula. Far from making it easier to write down, it introduces more room for error and confusion.
--- End quote ---
I'm sorry, but I think you are alone with this viewpoint.
There really is a reason to use different versions of letters in formulas. For example, \$v\$ can represent velocity while \$V\$ can represent volume, and both can appear in the same formula. If you tried to use the same letter \$v\$ for both, it would be hopelessly confusing. Similarly, we would typically have \$\epsilon\$ for an error, or for a small change, while \$E\$ would represent energy. Using symbols in a clear and consistent way like this aids communication and reduces ambiguity.
IanB:
--- Quote from: coppice on April 04, 2024, 05:49:13 pm ---Unicode is certainly a security nightmare. Between all the bugs in the character set (e.g. Chinese characters split into 2 separate ones that aren't actually different), the ability to express the same string in multiple ways, and other complexities, its nearly impossible to do a simple comparison between two pieces of text in Unicode. IBM produced a massive normalisation library in the late 90s, to try to make strings comparable, but its far from a complete answer, and isn't that widely used anyway.
--- End quote ---
Bear in mind we are talking about text documents and communication. How are there security problems arising from the text in a technical paper or a book?
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version