General

Useful links

September 9th, 2005

The latest Update from Cherryleaf landed in my inbox in the past couple of days and I finally got round to looking at it.

As usual it contained some interesting links - the prime ones being:

How to measure the effectiveness of a document

and

Cheat Sheets by UK web developer Dave Child

Potentially similar posts

Leave a comment

 

A smarter zipper

September 9th, 2005

I was looking for an application to unpack RAR files this morning and came across IZArc. I can recommend it as a smarter alternative to WinZIP.

Download it free from:

www.snapfiles.com/get/izarc.html

Leave a comment



Bruce Leroy Smith

September 2nd, 2005

Nothing to do with technical writing, just a very poignant blog entry:

http://aaronbsmith.org/archive/000104.html

Potentially similar posts

Leave a comment



CVS update codes

June 9th, 2005

When you run update, it outputs information about
what has happened. For example, if there was nothing to update it
will just say “Success, CVS operation completed”. Normally,
however, files have either changed locally (i.e. you’ve changed
something) or in the repository (i.e. someone else has changed something)
– or both. In this case, update outputs a
line for each file, with a prefix that tells you what has happened.
In TortoiseCVS the different operations also each have a different
colour, which helps you to see at a glance what has happened (e.g.
conflicts are in red).

The output codes are as follows:
(Note: Thanks to bioperl.org
from which this is derived.)

U  file

The local file has been updated to match the same file in the
CVS repository.

This is done for files that you haven’t changed but which
are not the most recent versions available in the repository.
It is also done if you update a directory and CVS finds a file
in the repository that is not in your working copy of the directory.

M  file
Your local copy of the file has been modified since the previous
update.

M can indicate one of two things about the local
copy of the file:

  1. You’ve change the file locally, but there were no modifications
    to the same file in the repository. In this case, CVS makes
    no changes to either file, but the local file and the file
    in the repository are now out of sync, so you will have to
    commit the changes you have made.
  2. There were modifications in the repository as well as in
    your local copy, but the changes in the repository file were
    successfully merged into the local file, without any conflicts.
C  file
A conflict was detected while trying to merge your changed version
of the file with the version in the repository. Generally this
happens because someone has committed changes to the repository
while you were working on a file. Usually you can avoid conflicts
by remembering to update a file before making any changes to it
– or if you are working on a file for a significant period
of time, save and update the file at regular intervals, and commit
any significant changes you make as often as you can (provided
you’re confident what you are committing is error free).
A  file

This signifies that you have done a CVS add, to mark that this
new local file needs to be added to the repository, but you have
not yet committed the file’s parent directory, which is what
actually adds the file to the repository. Basically, it’s
a warning telling you that you need to commit the parent directory
to add this file.

R  file

This signifies that you have done a CVS remove, to mark that
this local file should be removed from the repository, but you
have not yet committed the file’s parent directory, which
is what actually removes the file from the repository. Basically,
it’s a warning telling you that you need to commit the parent
directory to remove this file.

?  file

The file in your working copy of the directory does not exist
in the repository. Usually this means you’ve created a file
locally but not yet CVS added it and committed you change to the
directory. For example, if I edit a file called test.txt
using UltaEdit a test.txt.bak copy of the file
is always created in the directory. When I update the directory
this test.txt.bak file shows up in the output
as test.txt.bak.

Resolving conflicts

If you are using TortoiseCVS and there are conflicts, a dialog box
is displayed, listing the affected files:

resolve-conflicts-TortoiseCVS.gif

When you double click a file, you can choose the diff program you
want to use to examine the conflicts. A message box then tells you
the names of two temporary files (a “mine” file and a
“yours” file) and tells you to change the “yours”
file when resolving the conflicts:

resolve-conflicts2-TortoiseCVS.gif

Scroll down through the files, looking for highlighted conflicts:

resolve-conflicts3-TortoiseCVS.gif

At this point you’re supposed to edit the “yours”
file, correcting the conflicts, then save your changes. However,
I’ve never managed to do this because I get a message saying
I can’t save the file to the current location. When you close
the diff program TortoiseCVS asks you if you want to overwrite the
unsuccessfully merged file with the “yours” file you’ve
just edited.

The alternative solution (if, like me, you can’t use the above
method), and the method used by Windows CVS users and UNIX CVS users
alike, is just to open up the local file in a text editor and resolve
the conflicts there. But first you need to be prepared for what you
find in the file CVS has attempted to merge.

For example, I created a text file called test.txt,
containing:

start
one two three
end

I saved this file, then added and committed it to the repository.
Someone else then changed the file to:

start
one xxx three
end

They then updated and committed their change. I still had my local
copy open and I changed it to:

start
one changedlocally three
end

I then saved the file and updated it, at which point I got a conflict
message. When I open the local copy of the file, I see:

start
<<<<<<< test.txt
one changedlocally three
=======
one xxx three
>>>>>>> 1.2
end

You need to edit the bit between the line containing <<<<<<<
and the line containing >>>>>>>. The line containing
======= separates your version (above) from the repository version
(below). You need to manually reconcile these two versions and then
remove the marker lines. Once you have done this and saved the file,
you can commit the resolved version.

Incidentally, whenever there's a conflict, your original, unmodified,
file is saved as .#filename.version.
For example, if the file is called test.txt, the
original file is saved as .#test.txt.1.1 or .#test.txt.1.2,
etc. – depending on what version of the file you’re working
on.

See http://bioperl.org/userinfo/cvsdoc/commands.html
for more information about CVS commands.

Potentially similar posts

Leave a comment



login or log in or logon or log on?

June 5th, 2005

Is it correct to have a dialog box called "Login"? What about
a form field labelled "Login"? Or should it be "Logon"?

Some internet reading on the subject confirmed what I previously
thought was the case:

  • Login and logon are adjectives (or occasionally nouns) –
    not verbs. So writing "To login, enter your password ..."
    is incorrect, it should be "To log in ...".
  • Login tends to be used in UNIX, logon tends to be used in Windows.
  • Developers generally don't appreciate the difference and will
    use login, log in, logon and log on throughout an application without
    noticing.

    Note: Where I work, changes to the GUI of an application
    must be logged in Bugzilla
    and they are then prioritized (P1 meaning "Get this fixed
    NOW!!!", P4 meaning "In an ideal world, if we had a
    limitless amount of staff, we'd fix this"). A Bugzilla call
    asking for "To logon enter your username and password"
    to be changed to "To log in, enter your user name and password."
    would always, always be prioritised as a P4 bug.

Thomas Quine posted the following entry on GMANE back in September
2002:

The only thing I'd like to add to this discussion is a note about
the correct use of "log in" or "log on" (two
words) and "login" or "logon" (one word). Misuse
of these terms is rampant on the Web.

The rule is: one word, it's an adjective, two words, it's a verb.

So for instance, you can "log in" to a website, and
if you do so you are "logging in". If "login"
were a verb, then the act would be referred to as "loginning".

Login (one word) is an adjective. An adjective, you will recall,
is a word that modifies another word. Which is why it bugs me when
a see a field with the label "Login". My mind unconsciously
asks, "Login what? Login ID? Login username? Login password?
Log in using your email address, what? What word is Login modifying
here? What do they want me to enter in this field?" I have
come across websites that use this word to prompt for each of the
above. How's the poor visitor supposed to guess which one this
time?

My rule of thumb is, label a screen field with words that make
it clear what you want the visitor to type into that field.

This is not just being picky, because when we introduce even a
momentary confusion in the reader's mind, we introduce an obstacle,
and to me the key to information design is the removal of obstacles...

Source: http://article.gmane.org/gmane.comp.infodesign.general/499

Leave a comment



^ back to top ^

Page 9 of 14« FirstNewer7891011OlderLast »