Got more questions? Find advice on: ASP | SQL | XML | Windows
in Search
Welcome to RegexAdvice Sign in | Join | Help

Michael Ash's Regex Blog

Regex Musings

Grading the Guidelines.

Over at the Construction forum there is a sticky post which list the suggested posting guidelines for asking question on that forum. First let me say the guidelines were my idea however I didn't come up with them on my own. I had help. The idea sprang from the fact I was getting tired of repeatedly having to as each poster for the same bits of information. So the idea was to put these standard questions in a central highly visible location everyone could see. Where new poster could see what would be ask of them and add that information to their first post. The standard formula was


P1: Ask vague question

ME: Ask for more info

P1: Provide more info (maybe)

Me: Provide answer (maybe)


This being this be the best case but often didn't go this smoothly. Variations would be

ask for two piece of addition information get 1 (or none)

Answer question, Get response of “Oh I forgot about this” or “but what about this unmentioned case”

Where step two and three could repeat multiple times.


The point of the posting guidelines was to eliminated the extra back and for so you get down to

P1: Ask detailed quest

Me: Answer

P1: Thanks (maybe)


Now if I were the grade how successful the posting guidelines have been I have to give two grades.

Grade 1 (Posters asking questions): D-


Unsatisfactory passing. For the most part people asking the questions won't follow the posting guidelines. So in those cases nothing really changes. You still end up going back and forth until you get the information you need to properly answer the question. This is a waste of everyone's time, mostly for the person asking the question. For those asking the question, it just delays you getting an answer especially if the person responding is on a different sleep cycle than you. You ask a question, hours later someone on the other side of the planet starting their day responds “Need more information”. However you've signed off for the day. So the next day your respond but now the other person has signed off for the day (or weekend) So now instead of having an answer waiting on, you end up wait several more hours or days for an answer.


In some case getting the asked for information is like pulling teeth. I've always though it was strange how some people are so unwilling to help you, help them. Maybe it's because misery loves company and they want you to struggle with an answer in the same way they have been struggling with the problem. Sometime even after referring them to the guidelines people will respond but still not provide any of the information asked for. Sometimes they just rephrased their original query but add no new information or at least nothing satisfying any of the guideline request.


Some even get hostile about it. One person stated he couldn't answer any of the guideline questions because they were new to regex and didn't know anything about it. Now that wasn't really a valid excuse since the questions don't require any regex knowledge. The only one that even implies you have any regex knowledge is “Show what you have tried” which implies you have attempted to write a pattern. But if you haven't tried anything, you can reply “I haven't trying anything”


A lot of people with post questions with the mindset that you will answer while sitting in their lap. One of the reasons we are asking for this info is because we can not see what you are doing. We are not physically in the room with you. So statements like “I have a text file”, or “I'm using a text editor ...”

aren't helpful. We are not there, we can't see your file, we can't see your editor, which we may have never even heard of.


People will often post they are new to regex. Sometimes I think the feel this excuses them for needing to follow the guidelines. In fact it's just the opposite. They are the one who need to be following the guidelines the closest. We've all been in situations where we didn't understand something so much so we didn't even know what questions to ask. Or in this case what information is important and what is too much information. The guidelines cover the most crucial information needed. The more of this info provided the easier to answer your question becomes. The less provided the more difficult to answer your question becomes, irregardless if you are a regex noob or a vet.


The other thing, that I find annoying is people asking for help saying it's not important for the people answering to know what language they are using, or say it doesn't matter what language platform the use to supply an answer. Why do you think we would ask for unimportant information? The same is true for all the guideline questions. Do those asking the questions think these things were asked just because someone was nosy? Everything asked is so that you can get the best answer as quickly as possible. If you don't understand why the question matters, ask! But don't tell the people trying to help you what parts of the information they are asking for is important.




I said the the guidelines got two grades. The second if for people actually answering the questions. And for them the Guidelines get an A-(exceptional). The guidelines have been around a few years now even before some the active members providing the answers joined but in a lot of case I see them referring people to them with no prompting by myself. Not everyone answering reference the guidelines. Despite the things I just mentioned it is possible to provide an answer following none of the guidelines. It involves guessing and luck. Even if one ask a vague question, someone might be able to guess what they actually wanted and answer on the assumption and in some cases will be right. But that's mostly luck. Luck that they are familiar with the data you are working with. Just like above “sitting in your lap” people tend to assume that everyone is as familiar with the data or tools they are using as they themselves are. If you are lucky if that is true but nothing guarantees that. Plus a host of other assumptions have to be made and guessed correctly. Do you really want to rely on dumb luck every time?

 

A simple example

Q: I need make sure a sting contains only letters, what is the regex for that?

A: ^[a-zA-Z]*$


Assumptions made:

  1. By letters the question referred to letters of the English alphabet only

  2. Spaces are not allowed

Now that pattern may be exactly what the person asking wanted. But what if it wasn't. People asking these questions come from all over the world, English may not be their first language. The regex may not be intended to used against English only input.


The more common and actual request I see is when people say they need there regex to make “Special Characters”. Now maybe I was sick the day that they announced exactly which characters this term applies to and why they are special but when people ask for this there doesn't seem to be a consensus on which characters this includes. Now I could assume they meant any non-alphanumeric character, or even just non-alphanumeric characters on a keyboard. Of course this assumes we have the same keyboard layout. But often people only want a few of these characters. Which ones? Well that's part of the problem it varies. Even if just limited to the keyboard which characters are special or how many characters they are wanting differs by person.


One of the things that really tell that those answering question are good with the guidelines is when you see a person was answering questions or trying to that provided incomplete information get frustrated by the back and forth of such questions and start referring newer such post to the posting guidelines.


As new regex pros joined the board and start answering questions on a couple of occasions I have asked if they had any changes, additions or deletions to they thought should be added the to guidelines. The common answer is that the guidelines are good as is. The only real addition suggested were adding a Frequently Ask Question (FAQ) which while a good Idea since some question get asked repeatedly, does provide useful information for answering questions. Instead it would be more for preventing questions from being asked again. Only thing is the questions to vary slightly, like the afore mentioned “special characters” So the questions would likely still get asked especially considering the existing guidelines get ignored.


The only actual change from the first posting of the guidelines is the additional of a example of a badly asked question and template of a well asked question.


The actual guidelines can be found here http://regexadvice.com/forums/thread/60465.aspx but a quick summation of the 4 guidelines are


  1. What programming language/Application are you using?

  2. What are you trying to do?

  3. What have you tried to do?

  4. Show the actual text you are working with.


Now those who say they can't answer any of these question because they don't know regex, a probably are about to get fired from their job.

  1. What programming language/Application are you using?- If you are a programmer and you don't know the name language you are doing your work in you are definitely in the wrong job. If you are not a programmer or just using an application, how can you not know the name of the application you are using or planning to use. How did/will you launch it, or even find it on your computer? How do you know it will even allow you to use a regex?

  2. What are you trying to do? - If you can't answer this, well... You are not being asked about your process you are being asked for your goal.  If you don't know your own goal why are you even asking others for help?

  3. What have you tried to do? - Again either you tried something or didn't try anything.  If you have tried something please show it. If you haven't tried anything just say so.

  4. Show the actual text you are working with – Aside from security concerns which can be addressed, this basically copy and paste.

Why do we want to know this information?


  1. Regular Expressions are not uniform in all case. There are differences in syntax, features offered and even how features work that vary from host environment. And we aren't talking slight variations. In some cases the syntax of one regex engine is over 50% different from another. All pattern are not portable. Knowing the hosting environment makes it possible to know what is available regex-wise.

  2. Knowing the goal make giving an answer easier, if there is an answer. In some cases the goal may not be achievable with a regex or at all. Some solution require multiple patterns, some require programming code. Some goals are better suited for a non-regex solution

  3. Samples of what has been tried provide, insight to host environment, level of regex knowledge or knowledge of host environment. Code samples would show the language being used. Some attempts show misunderstanding of regex. Some attempts show syntax or logic errors in the host language code. Some people were close to getting the pattern right themselves. Others have had perfectly fine regex pattern but their host language code was wrong. So there is nothing with the regex that needs fixing there code needed to be corrected. Or even the pattern itself is fine but needs to be adjusted for the host language.

  4. This is very important but I have already covered that in another post.


Generally people incorrectly assume that if they make their question as minimal as possible it helps to make it easier to answer. As far as regex goes the opposite is true. The more details, the easier regex questions are to answer. Minimal regex questions often lead to more questions. Detailed regex questions often lead to quick answers.


Published Saturday, April 10, 2010 1:28 AM by mash
Anonymous comments are disabled