Timeline for How to properly choose between two equivalent implementations, and how to prove you picked the correct one? [closed]
Current License: CC BY-SA 4.0
21 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Sep 6 at 15:29 | comment | added | JimmyJames | @DocBrown I agree if the question is about precise meaning of 'prove'. But far too often, I've had the displeasure of talking to people who insist that because their subpar solution 'works' and therefore any critique of it is a matter of opinion. There are simply no practical, real-world criteria which justify wasting resources in order to protect someone's fragile ego. "there is no universally valid metrics for this" is the kind of thing I hear in such discussions. | |
| Sep 5 at 21:19 | comment | added | Doc Brown | ... we can "prove" a mathematical statement, or we can prove something in computer science about a theoretical algorithm. But in software engineering, we assess things like real-world implementations of certain algorithms, make a trade-offs, risk-assessments, do cost-benefit analysis and so on. | |
| Sep 5 at 21:11 | comment | added | Doc Brown | @JimmyJames: read my comment again, I never wrote all solutions to a problem are equivalent, or one cannot make a decision. My critics goes against the wording "prove (objectively)", as compared to "assess (by my own specific criteria)". | |
| S Sep 5 at 21:08 | history | closed |
gnat DavidT Doc Brown |
Needs more focus | |
| S Sep 5 at 21:08 | comment | added | Doc Brown | This question is similar to: If there are two ways of approaching a task, how should one choose between them?. If you believe it’s different, please edit the question, make it clear how it’s different and/or how the answers on that question are not helpful for your problem. | |
| Sep 5 at 21:06 | comment | added | JimmyJames | @DocBrown Two solutions can both solve the same problem but that doesn't make them equivalent. An incandescent light bulb and an LED may both produce the same illumination but have dramatically different energy usage. My bubble sort experience was someone proudly proclaiming that they would use bubble sort to sort items on the order of millions of items (on an ongoing basis) which were not expected to be in any particular starting order. The costs associated with that were not small. | |
| Sep 5 at 21:03 | comment | added | Doc Brown | ... when one of two implementations is "better" than the other, and that is IMHO a very wrong expectation. | |
| Sep 5 at 21:00 | comment | added | Doc Brown | @JimmyJames: "quick sort" is no objectively better than "bubble sort" (or vice versa). You need to define your criteria first and give them weight. Lets say my criteria here is *"I need a simple, easy to maintain sorting algorithm for 200 items at maximum which works in Excel VBA", and it needs to be just "fast enough" for my purpose". (I did not invent this to prove my point, I actually had this requirement a few years ago.) My point, however, is a different one: the wording of the question expresses an expectation about "mathematically strict, objective criteria" to decide ... | |
| Sep 5 at 20:38 | comment | added | JimmyJames | @DocBrown "You cannot really "prove" that you picked a "correct" design out of two when both designs technically work." I think you mean in general, right? I can sort a million items using bubble sort or I could use almost any other sort and, they both 'work' but I think we can objectively say which approach is better. This is actually a real technical discussion I had once and, yes, I do know there are some very specific cases where bubble sort isn't the worst option. | |
| Sep 5 at 20:21 | comment | added | Doc Brown | You cannot really "prove" that you picked a "correct" design out of two when both designs technically work. You can assess that a certain design fits better to certain requirements than another one (or not), but you have to defined the requirements, and you have to give them weights in an assessment, there is no universally valid metrics for this. | |
| Sep 5 at 20:15 | history | edited | Doc Brown | CC BY-SA 4.0 |
Nonsense in title fixed
|
| Sep 5 at 19:46 | comment | added | codemarc | I think this question is more about architectural decisions than tool selection. It's about structuring the system (e.g., monolithic vs. microservices architecture) rather than just choosing one tool over another (Supabase vs. microservices). My goal is to understand how best to organize the components of the system for future scalability, security, and performance. | |
| Sep 5 at 16:51 | comment | added | DavidT | It appears this whole post could be paraphrased as: "Should I use Supabase?" , if you already know SB and/or expect to do a lot of prototypes its probably worth starting new projects with it (i.e. fast start) whether it will meet long term needs is going to be based on project requirements - beyond that I am not sure what to add. | |
| Sep 5 at 16:51 | answer | added | Ewan | timeline score: 2 | |
| Sep 5 at 13:36 | comment | added | codemarc | How do i prove which one is correct. Are there any metrics that are commonly accepted as proof. If I do not know the what the usage max will be for this functionality. | |
| Sep 5 at 13:31 | review | Close votes | |||
| Sep 5 at 21:08 | |||||
| Sep 5 at 13:22 | answer | added | candied_orange | timeline score: 3 | |
| Sep 5 at 13:12 | comment | added | Greg Burghardt | I see two down-votes and no close votes. Can we have a little patience, please? The OP just posted 15 minutes ago. | |
| Sep 5 at 13:01 | comment | added | Greg Burghardt | The question title and body seem to be asking different questions. One might be on topic. The other one is definitely opinion-based. Can you clarify what you are asking? | |
| S Sep 5 at 12:56 | review | First questions | |||
| Sep 5 at 14:44 | |||||
| S Sep 5 at 12:56 | history | asked | codemarc | CC BY-SA 4.0 |