A couple of points:

• When I mention 'archive' in the title, I mean the Gmail 'Archive' feature - ie, remove the message from the inbox but keep it in 'All Mail'.
• You need to have IMAP enabled on your Gmail account.
• Don't let that '\\Deleted' stuff scare you - in Gmail's IMAP implementation, that removes the message from the inbox, but it doesn't actually delete it.

(... he said, ignoring the fact that he hadn't posted anything here in the past seven months.)

All of these services have a very similar feature set:

• A means to post a short entry, usually a textarea and a button
• A means to reply to an entry, again just a textarea and a button
• A list of entries and replies
• A means to "follow" other users, eg subscribe to their entries

I think that about covers the basics. Now, if you look around at Twitter, Pownce, Plurk, Rejaw etc, you might wonder why the textarea, button, and list provided by any one of these services is any more compelling than the same features on the other services. The answer is - it isn't.

Ping.FM addresses the box+button aspect of this by allowing you to update all of your services from a single location, and you can use RSS to aggregate timelines from the services in a single location, but you still need to jump around from one service to another to properly manage replies and subscriptions.

So - to the point... all of you microstatusblogamathings need to get together and come up with a way to do the following:

• Use some other service as a source for updates, and allow your service to be used in the same way. The textarea+button on your site is no better than the one on the other guy's site, so why should I be forced to use yours? Having my content on your site is what should matter to you - it shouldn't matter how it gets there. If you come up with some awesomely compelling feature that makes your textarea better than the others, fine... but have some basic standard that all of you guys can use to work together.
• Follow and reply to others, regardless of the service they choose as their "home" service, for the same reasons. Having this cross-site conversation visible on your site makes your site better, and it makes my life easier.

Now, do I think any of this will actually happen? Of course not. People have been making the same argument for interoperability in the context of instant messaging services for over a decade, and it still hasn't happened. But these new services are being built by a new generation of developers, so maybe they're a bit more open to this sort of suggestion than the giant corporations who control the IM services. Maybe it could happen.

Installation of Scrawl was straightforward, and after creating a few test entries with it, I was ready to migrate my news items over to Scrawl's "blog entry" content type. This was pretty easy, since Scrawl simply repurposes the News Item content type. First, I re-named the folder that contained my existing news items. Then, I created a new folder with the same ID as the old folder. Finally, I ran the following script (from my custom skin folder) to create Blog Entries with the same content and ID's as my old News Items:

from Products.CMFCore.utils import getToolByName

# get the catalog & workflow tools
pc = getToolByName(context, "portal_catalog")
wf = getToolByName(context, "portal_workflow")

# folder where the new Scrawl items will go
# change this for your own needs
nblog = context.nowhere.archive

# get News Items and create Blog Entries
entries = [e.getObject() for e in
            pc.searchResults(portal_type="News Item")]
for e in entries:

    # get info from the old object
    nid = e.id
    nkws = e.Subject()
    cdate = e.CreationDate()
    edate = e.getEffectiveDate()
    mdate = e.ModificationDate()
    ntitle = e.Title()
    ndescr = e.Description()
    ntext =  e.getText()
    ntype = e.Schema()['text'].getContentType(e)

    # create the new object
    nblog.invokeFactory(type_name = "Blog Entry", id=nid)
    item = nblog[nid]
    item.editMetadata(subject=nkws)
    item.setFormat(ntype)
    item.update(title=ntitle,
                text=ntext,
                description=ndescr,
                creation_date=cdate,
                effective_date=edate,
                )

    # my items are initially private, so two steps to publish
    wf.doActionFor(item, "show")
    wf.doActionFor(item, "publish")

    # set the dates on the new object to match the old object
    item.setCreationDate(cdate)
    item.setEffectiveDate(edate)
    item.setModificationDate(mdate)
    item.indexObject()
    print "created "+ nid

return printed

After running that, I had published blog entries that were copies of my old new items, with the same id's and in the same location as the old items, completely migrated with old URLs intact.

Feel free to copy and modify that script for your own needs - it's a simple example of how to migrate one content type over to another.

Edit: It just occurred to me that my script up there does not migrate comments. I only had a couple of comments on my old entries, so it's not a big deal for me, but it might be for you.

Not too crazy about this NuPlone theme, but it'll have to do for now.. stay tuned :)

Anyway - the point: steady income => more money => shiny new iMac => I'm no longer using Epiphany as my main browser, so I needed to cook up a way to quickly close a bunch of tabs in Safari. Safari already has a way to close all tabs except the current tab, so we just need a way to close tabs to the left or to the right of the current tab. AppleScript to the rescue!

First, the left tabs:

tell application "Safari"
    set cWindow to first item of windows
    set ctIndex to index of current tab of cWindow
    repeat ctIndex - 1 times
            close first item of tabs of cWindow
    end repeat
end tell
open location "x-launchbar:hide"

Then, the right:

tell application "Safari"
    set cWindow to first item of windows
    set ctIndex to index of current tab of cWindow
    repeat while (count of tabs of cWindow) > ctIndex
            close last item of tabs of cWindow
    end repeat
end tell
open location "x-launchbar:hide"

I save those as "Close Left Tabs" and "Close Right Tabs" in ~/Library/Scripts. When I need to use them, I just bang a couple of keys to bring up the very-excellent LaunchBar [1], then 'CL' or 'CR' to run whichever one I need.

So, this is great for me and Safari, but it doesn't bode well for the similar work I did for Epiphany. There's a problem with that one, I think related to the disconnecting of GTK signals. The symptom is that if you're using that extension, Epiphany won't cleanly exit after you've closed all of its windows. And the real problem is that since I'm not in GNOME most of the time these days, I'm really not feeling an urgent need to fix it, which sort of sucks if you happen to be using this extension with Epiphany.

So, if you're one of those two people... feel free to hack on it ;)

I downloaded the deb package, installed it, bought an album, and was immediately presented with a fun dialog similar to this:

JOY. What could have been a seamless, 1-click music purchase was changed into a clumsy series of download/locate/click steps. I can't really blame Amazon for this - it's Epiphany making the safest decision when presented a brand new MIME type, which isn't such a bad thing in and of itself.

(However, GNOME will, without a peep, happily allow me to unleash this file's document-damaging, privacy-invading wrath once I've downloaded it.)

What is a bad thing is that there's no UI in Epiphany to let me decide what's safe and what isn't. Isn't setting up helper apps for file types a pretty common thing? You can't even fix this as a regular user - to add a new safe MIME type to Epiphany, you have to add it to the "safe" section in /etc/gnome/epiphany/mime-types-permissions.xml, which of course requires you to be root.

(In this case, the MIME type for Amazon's AMZ files is <mime-type type="audio/x-amzxml"/>)

GNOME is my desktop environment of choice, and I admire the GNOME project's goal of simplicity, but forcing a user to muck around with XML inside of a system configuration file seems to be contrary to that goal. When you dumb things down to the point where you're removing the means to change common settings, you're going a bit too far.

Like I said, it's nice to have options. :)

If you'd like to try this out, download Close Other Tabs, extract it, open a terminal in the extracted directory, and run 'make install' to copy the extension to your Epiphany settings. Once youve done that, you can activate it under Tools > Extensions in Epiphany. If you don't have 'make' installed, you can manually copy close_other_tabs.* into ~/.gnome2/.epiphany/extensions.

I've still got some work to do on it, primarily on UI niceties like disabling the menu items when only one tab is open, etc. I also need to put some credits in there - the layout of the README, the Makefile, etc all begain life in a different extension. :)

(Edit: Done!)

Anyway, if you try this out, please let me know what you think!

2008-03-07 I'm not sure why, but this extension is causing Epiphany to leak memory and not exit cleanly. Something to do with disconnecting signals I think, but I'm not sure what I've done wrong. :/

I've received a few emails asking when I might push out an updated version for Plone 3, and the answer is always "when I have the time". But after reading Martin Aspeli's latest post about Plone 3 portlets, I can change that answer to "never" and not feel bad about it...

"plone.portlet.collection allows you to point a portlet at a Smart Folder to get any type of listing. It's not 100% done yet, so it doesn't have a release, but it's nearly there."

That's pretty much what SmartPortlets did, so I suppose I'm off the hook!

I don't think I'm going to stay with this particular setup for my blog... I rather like having full entries listed on the blog front page. I'll probably end up using Quills, but we'll see.

At any rate, the dust will settle around here sooner or later.. sorry for the inconvenience!

I looked at two advertising systems for Plone:

• KBannerAd
• PromoEngine

KBannerAd

KBannerAd is an older product from the Plone 2.1 era, but I have tested it with Plone 2.5 and it does work without any problems, although there are a few UI glitches.

For creating ads, KBannerAd is simple to use. You create an ad using a tool in the Site Setup area. The ad object has fields for:

• an image (can be gif, jpg, png, or flash; text ads are not supported)
• a destination URL (either internal or external)
• a rotation weight
• effective and end dates
• location of the ad (ie, which portlet or banner slot)
• maximum clicks

Ad rotation is essentially random, but weighted - if ad A has a weight of 80 and ad B has a weight of 20, then over a few hundred page views, ad A will have been displayed approximately 80% of the time.

The lifetime of an ad is determined by the effective and end dates, and by the maximum clicks (whichever comes first). The max clicks allows the site owner to sell ads on a cost-per-click basis (meaning, the number of times the ad is actually clicked). The ad tool in Site Setup shows how many times each ad has been clicked.

Ads are controlled by workflow, so if an ad is not in a state accessible to the user, the ad will not be displayed.

PromoEngine

PromoEngine is a newer product, under active development using the latest Plone development techniques. It is only at version 0.5.1, but I have not encountered any problems with it in my tests.

For creating ads, PromoEngine is a bit more complex than KBannerAd, but with that complexity comes a bit more power in the ability to organize and display ads. PromoEngine uses a tool in Site Setup - you add folders under that tool to organize your ads. To each of these folders you can add subfolders, images, and ad objects. An ad object has the following fields:

• Reference field to associate ads with specific content
• The portlet slot (or banner location) where the ad should be displayed
• A Kupu editor to create the ad content - basically anything you can build in Kupu can be used for an ad, so you could have images, Flash, text, links, whatever. A catch here is that, for example, if the site owner is building a banner ad, they must remember to link the image to the destination URL. This is slightly more difficult than KBannerAd, which automatically wraps the banner ad with the destination URL.

The reference field is a good feature - it will allow you to run specific ads with specific content items (articles, etc).

PromoEngine does not have the max clicks and rotation weighting that KBannerAd has. The life of an ad is determined strictly by workflow and expiration date, and rotation is completely random. Weight is a planned feature, but is not yet implemented.

Common Shortcomings

• Neither system will do per-impression rotation.
• Other than the click counts provided by KBannerAd, neither system provides any type of statistics reporting.
• Neither system has the ability to manage the actual sale of ads
• Neither system can create a sized pop-up window when an ad is clicked (although this might be possible with PromoEngine, with some tweaks to Kupu).
• Since both systems serve ads directly in the HTML of a page, CacheFu page caching will prevent proper rotation of the ads.

Feature Comparison

Recommendation

Based on all of this, both of my clients have decided to go with PromoEngine, primarily because:

• It is a newer product, with active development.
• It allows the organization of ads into folders and subfolders in a flexible manner - you could organize your ads (and their associated images and other resources) by client, campaign, whatever.
• Rich-text (Kupu) creation of ads allows for more creativity in your ad design, and allows for text-only ads.
• Allows specific ads to be associated with specific content items.

Stuff I'd like to see

• The simple image+URL ad types in KBannerAd are nice, would like to see those in PromoEngine.
• A wrapper around a PromoEngine text ad, to make the whole thing clickable.
• A way to designate an ad destination as a sized pop-up window.
• Full reporting for views and click-throughs.

If work ever slows down, I'll try to find the time to contribute some of that. But working with good open-source projects is a double-edged sword - if you make the right connections and land some good clients, you'll find yourself too busy to contribute back to the projects!