2012/6/28 Armin Le Grand <[email protected]>: > Rob Weir <[email protected]> wrote: >> On Thu, Jun 28, 2012 at 6:57 AM, Armin Le Grand <[email protected]> >> wrote: >>> Hi Kevin, >>> >>> Kevin Grignon <[email protected]> wrote: >>>> KG01-see comments inline. >>>> >>>> On Thursday, June 28, 2012, Dan wrote: >>>> >>>>> Srinivasulu Bhattaram wrote: >>>>> >>>>>> I have a text file, which has one space before every line >>>>>> How to remove this using search and replace option >>>>>> For carriage return or <Enter>, I learnt that we hould use "\n" (quotes >>>>>> not included) and keep Regular Expressions on. >>>>>> But I do not know what is code for space >>>>> >>>>> >>>> KG01 - Why is this so complicated. Such system language hides the power of >>>> the tool. The tool should understand natural language inputs. >>> >>> I already thought the same. Regular expressions are mighty, but only >>> (guessed) 5% of the users understand it (as with many mighty things). My >>> first idea was to add buttons which add the 'coded' entries tothe >>> expression, e.g.: Button 'Space' adds '\n' when pressed to the text field. >>> Same for 'Start of Line' adding '^'. Much more to be defined, may someone >>> knowing regular expressions well list more of them. >>> Another way would be to allow tokens like <Space> which get translated to >>> '\n' internally before using the expression, but then the user would again >>> have to remember multiple tokens and how they are spelled (need to be >>> valid, possible errors which need to be handled). >>> Maybe a mix of both, press button 'Space' adds the (human readable) token >>> <Space> which itself is not editable (represents a single character). >>> Hmmm... >>> >> >> Have you seen what the SlickEdit coding editor has? Here is their >> search dialog: >> >> http://people.apache.org/~robweir/slickedit-re-dialog.png >> >> So you can click a button to pop out a list of common RE patterns. > > Interesting. What is inserted into the text field, then? Some > human-readable tokens, or the RE ones? > >> But this is more of an aid to someone who already understands RE's. >> I'm not sure we can do much to make this conceptually much easier for >> end users, since regular expressions have a certain amount of >> essential complexity. Or would it make sense to support more than one >> RE variation, maybe full RE, but also DOS-style with only * and ? >> qualifiers? > > That is what MSDEV (Developer studio) does. It allows to choose between > 'Wildcarts' and 'Regular expression', where the first one accepts simple > '*' placeholders. An alternative for users, but limited. Maybe a good > compromize... > >> -Rob
If you have two choices and you pick both, it's not a compromise, is it? Kind regards Johnny Rosenberg ジョニー・ローゼンバーグ --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
