As AndalayBay suggested, I've opened up a separate thread for discussion of changes to the BOSSlog to stop it filling up the main BOSS thread and taking attention away from mod reports, and to hopefully get some more input from BOSS users that don't necessarily follow the main thread.
http://dl.dropbox.com/u/17043363/BOSSlog165.html (the latest release).
http://dl.dropbox.com/u/17043363/BOSSlog.html, which will be released at some point, hopefully sooner rather than later.
The link for the v1.7 BOSSlog will always point to the latest iteration of the log.
I'd like to hear your thoughts on the new format and am open to any feedback/proposals you might have.
BOSS v1.7 is actually in a state where it was ready for release a couple of months ago (waiting for Wrye Bash 292), and all these log changes were scheduled for v1.8. Since they're all HTML/CSS/Javascript-based and so don't impact BOSS's execution code they're safe enough that I can make the changes without having to go through another testing cycle, allowing me to make more improvements before release. I'm not making any more changes that require execution code changes. As such, not everything I agree to put in may make it into v1.7, but if I say something will go in, it will be in for v1.8.
Needing Discussion
- Having a summary at the top of the BOSSlog. Opinions? What should go in it?
- Colour-coding of message types besides warnings and errors? Should all message types have their prefixes colour-coded? If so, what colours?
- Do messages stand out from their mods enough?
- Is the log easily readable?
- Do all the log's filters and collapsible sections work in your web browser? (Confirmed working for Firefox 4, Google Chrome 11, Opera 11, Internet Explorer 9)
Some Reasons For The Changes
Filters
These allow the user to pick and choose what is visible in their BOSSlog, so they can cut out stuff they're not interested in, making it easier to see the important stuff. I don't think there's any downsides to them being there, so there's not much more to say.
Collapse-able Log Sections
Those minus signs to the left of the BOSSlog section sub-headings mean that the section is collapse-able, so clicking on the sub-heading will hide its section, and the minus will turn to a plus. This is again a feature for hiding information the user is not interested in, and for saving the user from more scrolling if, for instance, they only want to see what mods were unrecognised.
Colour-coded messages
Green = OK, Orange/Amber = Warning but not serious error, Red = Serious Error. Previously only Error messages were coloured (in red), but the addition of the green colouring sends a clear signal to users that they don't really need to read the information. The message types and their colours can be seen listed under Oblivion.esm's entry in the v1.7 BOSSlog.
OBSE & OBSE Plugin Checksums
Not a format change, but people might ask questions about this new section, this is only displayed when you run BOSS with the --crc-display, -c flag set. It's sole purpose is for people to see what the CRCs for their OBSE plugins are, as this information can be used in the masterlist. So it's primarily a BOSS team tool, but visible to anyone who wants to see it. Not much more to say.
Ghosted, Version and Checksum Labels
These were all given background colours to distinguish them from the rest of the text. The Checksum information is new, it is only displayed if BOSS is run with the --crc-display, -c flag set. People said that the background colours made the labels stand out too much, so I turned their background colours into their text colour. The result is that they are still distinguishable but not in-your-face. Although the labels might be too light to read for some people now.
Non-Bolded Mod Names
The old BOSSlog used bold to highlight mods with messages. Some people complained about the formatting inconsistency with the rest of the mod names, and I agreed. I had set all the mods to be bolded, but this gave emphasis to the wrong parts of the BOSSlog, as the messages are important, not the mods, since messages might have to be acted on, but mods never do. I still needed some way to differentiate between the mods and their messages besides the indent, and so I italicised the mod names. Then those with visual impairments complained, so I got rid of the italics too, and just increased the amount of indent slightly. Hopefully that should be enough.
More Spacing
The introduction of colour to the BOSSlog meant that more spacing was required to stop the colour from 'bleeding out' into the surrounding text. In addition, I feel that the old BOSSlog didn't give enough spacing between sections of text, which would lead to the text being harder to read. I actually slightly decreased space between BOSSlog sections, but increased the space below each mod's entry, visually weighting a mod's messages towards the preceding parent mod rather than the neutral spacing used previously.
Removal of "Unknown Mod File: " Text
The text was superflous given that the mods were appearing in the Unrecognised Plugins section, so I got rid of it. Less text to read without a loss of meaning is a good thing in my book.
Plugin Numbers
This is another new section that gives a few numbers relating to the user's mods. It's been proposed that this be moved to the top of the BOSSlog and turned into a fully-fledged summary section, which is something I'm interested in discussing.
Answered Questions
Q: What about being able to set defaults for the BOSSlog filters? Will we be able to do that?
A: Not for v1.7. I'm considering introducing an optional config file that will allow users to set defaults for various parameters for v1.8.
Q: What about the BOSSlog being a diff of what's changed since the last time BOSS was run so people don't need to read it if there was no change?
A: The format of a diff is generally counter-intuitive to the average user, I think, so it is unlikely. A simple message saying whether the output has changed at all since the last time BOSS was run is much more feasible and is something I can do. It won't happen for v1.7 though, because it would require more processing code changes than I'm happy doing at this stage in the release cycle.