Recently I found a link from the Rss Bandit referrer stats that points to the Microsoft Technet Security Site: there you can read about RSS, but more important you find a link to the security bulletin feed! Subscribed.

Yes, it is now also available via my favorite magazine! Have a look here (german).

Geeks that know about RSS version(s) may already have thought about what should happen to the post date of an yet published item: don't think more about it - keep it as it is!

Why? This provide some kind of "permanent" web page(s) - the same thing as we want to have with the permaLinks of posts.

And here is the other reason. Ian Hickson posted an update of the state of the WHAT Working Group. BitWorking referred to it in his post 3270 Redux from June, 6th 2004 as I did just some words before. Now a RSS Bandit user complained about, Ian's feed is not displayed/processed correctly in the list of items. The item post date in the feed source is of 2004-06-29T16:26:32+00:00, but it displays June, 6th 2004! How comes?

This behavior is not a failure in feed parsing/processing. We read the date(s) in correctly, I double verified that. But if you look carefully you will note there are only a few (or at least one) item(s) where the date is "wrong". E.g. "State of WHAT" as of June, 29 2004, but only if you also have subscribed to the BitWorking feed.

I assume here, especially this item is a edited post (as of date above) after it was published (before the date you see in Bandit for that item)^{(see correction note below)}.

As you know we display item relations in threads. So our Relation Cosmos implementation is so smart, it corrects the item date(s) if it detects "invalid" relations from one item to another. So it does for Ian's post with an referrer item posted at June, 6th 2004: this mean, Ian's post must be already published that time, because the referrer has the link to it. So the date of that particular post is "corrected" back to be a few seconds BEFORE the item post date of the referrer post. The time line of posts is now correct again.

Conclusion: leave the item post date unchanged, if you really have to "correct" the posted content of an item! Or a smart software does it for you! (Or even cannot display a time line of posted items in a thread or any other visual representation like ThreadArcs). Or: one can even "migrate" to Atom, that maintains also a changeDate of items if you really need it...

Edited correction: As Dare Obasanjo tolds me, the story was BitWorking edited an old(er) post to contain a link to the newer one (Ian's). But left the post date unchanged (as I suggested). How common ist that? What do you think? As it is somewhat unclear what item post date we had to correct, we will left it untouched in the future... :~

Today I stumbled over this little balloon window popping up as I pressed Ctrl-C on a simple textbox with password char set to '*':

BTW: who wonders about the green progressbar - it is a password quality meter you may know from Mozilla browser...

Just watched the WWDC 2004 keynote video [via Steve's blog], then read Omar's comment: he is so right about Tiger (ships in 2005), and Longhorn takes too long to get more focus for now. When I have to buy a new machine for home usage, it will be definitly a Mac, with preferred Tiger installed. For this I can wait the year. Maybe with a new 23" display (the new 30" one is far too expensive for me, but really cool! With the dual-dual DVI cable and 4.1 million pixel resolution...)

Wouldn't a port of mono to Mac OS X (Tiger) with a System.Windows.Forms implementation using the Mac widgets be a cool project?

Lockergnome: Penguins sited in Munich. In reality they will use then a linux distribution at their desktop and run MS Terminal Server session(s) to get their every day applications (e.g. Access and Excel VBA apps). Differences:

1. Much slower than before (via terminal server)
2. If TS is down, no one can work, and: how expensive are TS licenses?
3. IT division is down for all the requests they get from their users to impl. just small additions to the new app running now at linux (or they apply it in Excel VBA to be in time; see point 1).
4. Linux and the main stream apps like Office or Image processing have  community support, but government software?

Nothing against Unix/Linux (on servers).

Doug Lawty posted at Weblogs@ ASP.NET a very useful tip: don't press Alt-PrtScn to get a clipboard copy of the content of a messagebox (as a image), instead use simply Ctrl-C (just text)! It also works for .NET messageboxs like the Rss Bandit About box:

---------------------------
About RSS Bandit
---------------------------
RSS Bandit 1.2.0.114 written by
* Dare Obasanjo Obasanjo (www.25hoursaday.com/weblog/)
* Torsten Rendelmann (TorstenR, www.rendelmann.info/blog/)
* and all the active members of RSS Bandit community.

Credits:
* Mike Krueger (#ZipLib)
* Chris Lovett (SgmlReader)
* SandBar Copyright (c) 2004 by Tim Dawson, http://www.divil.co.uk/net/
* Portions Copyright ©2002-2004 The Genghis Group (www.genghisgroup.com)
* sourceforge.net team (Project hosting)
---------------------------
OK
---------------------------

But maybe that's little bit too much to just get the version number copied

Yesterday I read a post from Chris Pirillo about a RSS Reader Security Check Tool. So I tested RSS Bandit and here are the good news: all tests succeeds without a security warning (v1.2.0.114). There is one exception: if you have it configured to display the link taget site of feed items that does not provide a description, then frames are allowed (if we would not allow frames, some news sites looks very wired). So the iframe test results in a security warning. Does anyone know a way to control iframes independent of normal frames within IE? Means: switch the download on/off independent of each other?

Some of the Rss Bandit users reported an issue with network connections using a proxy that require authentication (forum thread http://www.rssbandit.org/forum/topic.asp?TOPIC_ID=260). All of them should try the provided binaries zip (link available at http://www.rssbandit.org/ow.asp?YetImplementedFeatures) and please report any further existing issues at the above forum thread.

Today I stumbled over a framework v1.1 issue related to the wellknown System.Collections.Specialized.NameValueCollection. I used it in a public interface as a return parameter and somewhere again as method parameter within a assembly marked with the attribute CLSCompliant(true). The MSDN docs are silent about CLS compliance in that case and I figured out, it is not compliant! Got compile errors like this:

ADSIAuthenticationProvider.cs(31,37): error CS3001: Argument type 'PROCOS AG.Components.Authentication.Shared.IAuthenticationConfiguration' is not CLS-compliant

or

And still very busy at work. Not much free time left to get more progress at Rss Bandit, but recently we get the bugfix release 114 ready for download.

Just read this at Early Adopter's:

"Every line of code you write today (and have to maintain) is 4 lines of new code that you don't get to write in the future. Write your code wisely." [80% of the cost comes after RTM]

He is soo right. I should have made the quote bold and 40pt size...