Talk:Sorting algorithm: Difference between revisions

Content deleted Content added
Example stable sort vs. FERPA
Line 806:
:I agree with "caching effects on modern (or not so modern) processors can greatly affect the speed of some memory access patterns". "(...) and that could be significant in some sort algorithm comparisons", yes. For instance, a sorting algorithm which works on the whole input from the start (such as bubblesort) is less cache-aware than an algorithm such as mergesort which can focus on local solutions, which tend to be contiguous in memory, before going over a larger slice of whole input. [[User:BernardoSulzbach|BernardoSulzbach]] ([[User talk:BernardoSulzbach|talk]]) 22:28, 4 September 2019 (UTC)
::One example of this is on a processor with an 8 way cache, a 4 way merge sort (without heap, just an if/else tree to determine smallest element), there are 4 cache lines for the 4 runs being merged, and a 5th cache line for the merged output. The end result in my testing is that 4-way merge sort is about 15% faster than 2-way merge sort. [[User:Rcgldr|Rcgldr]] ([[User talk:Rcgldr|talk]]) 02:53, 6 September 2019 (UTC)
 
 
== Example stable sort vs. FERPA ==
 
The dynamic web page in the second paragraph implies (at least to me) a '''Publicly accessible''' web page, whose content runs afoul of [[FERPA]].
 
The second paragraph of the Stability section starts with:
 
Stability is important for the following reason: say, if the data is sorted first by '''student name''', in some cases, dynamically on the webpage, and now the data is again sorted by which '''class section''' they are in.
 
A webpage containing '''student name''' and corresponding '''class section''' data would at first glance appear to be "''''directory information'''" but is actually an '''education record'''. Therefore if the web page is accessible by others than those with a defined, institutional, education purpose, the web page would be in violation of [[FERPA]].
 
My feeling is those learning these topics should not be creating things, that, if used with real-life data, would be in violation of the law.
 
{{Cite web|url=https://studentprivacy.ed.gov/faq/what-education-record|title=What is an education record? {{!}} Protecting Student Privacy|last=|first=|date=|website=studentprivacy.ed.gov|publisher=US Department of Education|via=https://studentprivacy.ed.gov/frequently-asked-questions|url-status=live|archive-url=https://web.archive.org/web/20181225033723/https://studentprivacy.ed.gov/faq/what-education-record|archive-date=Christmas 2018|access-date=26 February 2020|quote=[...]records include but are not limited to grades, transcripts, class lists, student course schedules, health records (at the K-12 level), student financial information (at the post secondary level), and student discipline files. [...]}}
 
 
{{Cite web|url=https://www.registrar.psu.edu/staff/ferpa_tutorial/ferpa_directory_info.cfm|title=FERPA Tutorial - Directory Information{{!}}When is Directory Information Not Really Directory Information?|publisher=Office of The University Registrar - [[Penn State]]|url-status=live|archive-url=|archive-date=|access-date=26 February 2020|quote=It is important to also understand the concept of "implicit disclosure." An implicit disclosure may occur when a list consists only of directory information but the list itself by definition reveals non-directory information. For example, a list of names and email addresses of all students who have a particular grade-point average reveals the students' GPAs. Likewise, a class list containing names and email addresses of the students reveals class enrollments. '''Since neither''' grade-point average nor '''class enrollment''' are '''directory items''', releasing these lists without prior consent of the students constitutes a FERPA violation.}}
 
--[[User:Lent|Lent]] ([[User talk:Lent|talk]]) 08:55, 26 February 2020 (UTC)