Drupal reminds me of Google Page Creator

My now moldy vintage 1994 personal web site [1] was created using Microsoft FrontPage, especially FrontPage 98. That was an impressive piece of software for working with document-oriented web sites, unfortunately there's nothing like it around today. The closest thing to FrontPage today, curiously, is Microsoft's Sharepoint 2007 wiki.

FrontPage 98 still runs. In fact, I recently connected it to the XP SP2+ IIS-based personal web server. Unfortunately, it doesn't run on my current platform of OS X (save in emuliation of course).

So I've been looking for replacements, albeit sluggishly. Since even Dreamweaver seems to be a relic of another age I thought I'd look at Drupal, an open source "content management system" -- meaning a web app for authoring web pages.

It was easy to add to one of my DreamHost domains - it's a free one click install there. I had it up and running in (really) less than a minute.

Alas, even though DH claims that they'll keep it current it's already one big security update behind (so no links from here!).

Still, I was able to play with it. It reminds me of the late midly lamented Google Page Creator -- except that there's no rich text editor. In fact, the Firefox editing window I'm typing this into is far more powerful than a Drupal editing window.

I'll keep poking around, but I sure do miss FrontPage ...

[1] In one form or another I've maintained a continuous Internet presence since 1994. Holy cow.

Transferring a Google Apps eNom domain: Awkward!

[Updated from the original, it's not as bad as I first thought.]

I'm in the midst of correspondence with several eNom support people. The upshot is that I recommend against the "easy" default path of registering a domain through the Google Apps process.

Instead I recommend working with DreamHost (my most recent blurb on them, with discount code). They will provide a full service registrar function and any desire mix of Google Apps and standard web and web app services.

It all started when I registered a Google Apps domain for a specific project. I then realized I needed a true web server with the domain, rather than Google's moribund and dysfunctional "Sites" service.

No problem, I've done something this before, though this time I simply chose the DreamHost web+Google Apps option. I then changed the DNS settings at eNom and everything just worked (DNS propagation meant a full switch took about 10 hours, but there was no service discontinuity). I had the standard Google apps and my domain name now pointed to a web server.

[Update: Normally when you change DNS settings at a registrar I think you can enter either the IP address or the readable name (ex: NS1.dreamhost.com). Not so with eNom. The IP address won't work. Sheesh.]

Then it occurred to me that there was no obvious way to unlock the eNom managed domain name and request a domain transfer.

So I asked customer support.

I must say they answered quickly, but the answers were very confusing. I think after parsing them out (updated here, I had this wrong before) they can be moved after 60 days post-registration but the process is pretty manual and documentation is hard to find:

This is the best official documentation I can find ...

Unlock domain (GoDaddy and eNom) - Google Apps Help

...eNom

To unlock your domain with eNom, please contact the eNom support team to request this action. You can reach eNom.com support via email at googleclients@enom.com, or via phone at 425-974-4623. These channels are dedicated to Google Apps administrators who registered a domain with eNom during the signup process...

Messy. I think I'll gradually consolidate my eNom domains with DreamHost. (To be clear, I don't speculate in domains -- I use 'em all!)

Update 3/8/09:

eNom does have a process for transferring domains, but it's not publicly documented.

The following is translated from what I was sent by eNom customer service. They're very responsive but I'm pretty sure they aren't native English speakers. If you email keep things very short and simple. I think the phone number option is preferable.

... in order to transfer the domain, it needs to be 60 days after registration...

... This isn't an eNom-specific rule - this is a rule for all registrars set by ICANN.

If a domain name has been registered for more than 60 days, email with a request. Include the domain access password for verification. (This isn't your Google pw -- NEVER give that out.)

The domain access password business can be tricky.

Remember that in the Google Apps domain admin page there's an "advanced DNS settings" link that will display your eNom domain access password (and a link to the eNom admin page, but that's not relevant here.)

If you haven't changed the eNom domain access password Google generated you just need to include this password in your email.

If you've changed the original pw using the eNom admin page and haven't lost your new password, that's the one you put in the email (frankly I suggest phoning instead).

If you've lost your revised pw here's where you're reminded how critical it is to retain control of your email accounts (read this as a reminder!)

simply put in your domain name at the access login screen and then click on the "forgot password" link and it will be emailed to you at the address on file with Google for your domain name.

Yikes. You really don't want to lose control of your Gmail account.

Once you've submitted the unlock request, in a day or so your domain will be unlocked. At that point you can initiate a domain transfer request from DreamHost (or any other registrar).

I don't know what happens next, but I suspect eNom sends you an email when they get the domain transfer request and you have to validate that. Either that or the domain transfer. Or something. I think I'll do this sometime in the next few weeks, so wait for the next update ...

Why you want XP, not Vista, for use with an OS X VM

Great series of tests by MacTech, excerpted by Apple Insider:

AppleInsider | Shootout: Parallels outperforms VMware Fusion in many tests

... 3D and HD Graphics Tests

XP: Smoothly played both 720p and 1080p videos in both environments.
Vista: Couldn't play 720p at all in Parallels. VMware Fusion stuttered on every machine except the Mac Pro. Given 720p results, MacTech didn't bother trying 1080p...

Yikes! Rules out Vista for me. You really want XP inside these VMs.

Parallels was significantly (20% range) faster than Fusion, but for me speed is less critical than reliability and stability. They didn't seem to test that.

Streaming video over the net - Quicktime, Ustream.tv and Watershed

We've been making extensive use of Google Video Chat for corporate communications (and with my aged mother, but that's a future post), but it's strictly point-to-point.

We need to share a video stream (audio not needed) from site meetings to remote users. This is remarkably hard to do.

You might not think this is a challenge. You might recall a 1990s fad of using a desktop webcam to share daily tedium -- or webcams that broadcast traffic. Or you might think of a large industry that specializes in "recreational" webcam use (an industry that just about killed the quality desktop webcam).

Alas, it would be most unwise, not to mention unsavory, to use those recreational services for corporate video broadcast. Besides, we actually want image clarity.

I haven't been able to find many options other than the high end professional services.

The one thing I've come across is the combination of Apple's recently resurrected QuickTime Broadcaster for OS X (compresses video input) and Apple's somewhat quiescent QuickTime streaming server.

Apple's free QuickTime Broadcaster for OS X (FAQ) will support firewire video capture, such as from a Canon Camcorder as well as iSight input. It can only output to a single destination however, such as a (Windows/Mac) QuickTime Player or (more importantly) OS X Streaming Server. (Yes, the name is misleading. Also the documentation is obsolete, iSight no longer exists but it now works with any 10.5 video source).

In theory it works with both Intel and PPC machines, but my G5 iMac couldn't compress the high quality video output of my Logitech Vision Pro webcam fast enough.

To do real broadcasting you're supposed to stream the output to a QuickTime streaming server (part of OS X Server, $400) or a multicast network. (This discussion is useful).

It turns out that DreamHost, a well regarded web hosting service, provides the open source version of QuickTime streaming server -- the Darwin QuickTime Streaming Server. Live streaming (broadcast a meeting) is not officially supported, but it works. The configuration looks like this ...

• Local OS X laptop provides live feed (OS X Broadcaster) to Darwin Streaming Server
• Darwin Streaming Server provides on demand stream
• Users access stream from a specially configured web page that embeds QuickTime call.

I've played with this configuration briefly, but there's very little material on the web about it. That makes me wonder if there's any way to make it really work (Apple is very quiet, for example). However I found IAMedia really had used DreamHost's streaming video. They've prepared a nice tutorial of how to make it all work, including how to embed the stream in a webpage.

Problem is, they've run into quality of service issues with DreamHost. So they've recently switched to ustream.tv -- a ad-funded startup specializing in personal broadcasting.

Alas, ustream.tv isn't very corporate, though it's not as off-base as the "recreational" services of old.

Happily, ustream.tv sells a private label service called "Watershed".

... Watershed is Ustream's self-serve platform for live, interactive video. Flexible for everyone, Watershed offers plug-and-play as well as robust API integration solutions. Organizations both small and large can customize Watershed to meet their specific needs and build global communities around shared live experiences....

Watershed charges $1 an hour/user for pay-as-you-go pricing.

That's about right for my corporation ...

Update: Watershed isn't super trivial to setup, but by the standards of video streaming it's very simple. I created the two web pages (broadcast and viewer) on one of my servers and stuck the embedded code in. Worked pretty well. Cost for our use would be about $50 to $100 monthly, so it looks like something I can justify.

So I was wondering, where the heck was Watershed all the time I've been looking for an affordable corporate video broadcast solution?! Turns out they launched 2 weeks ago. They're probably not even advertising yet.

Update 3/6/09: A few cautionary notes on Watershed

• I don't see an automated way to discontinue an account. I do like to see that.
• They don't provide any information on which credit card you're billing billed against
• The "Support" link doesn't have any link to contact support (there is a separate contact link)
• When I tried it this morning it was broken.

Update 3/24/09: After my initial testing I was never able to get it to work. Tech support was responsive, but it didn't clear up the trouble I was having. I decided to step back and wait until there are more players in this market. Then I discovered there was no way to remove my account information...

The sorry state of OS X device drivers

I've written before about the lousy quality of OS X scanner drivers. Today Scott Gruby, a vendor of OS X software, provides some details on miserable Epson drivers and makes a limited recommendation:

Scott Gruby’s Blog -- Accepting responsibility for bugs

... The only scanners I recommend are the Fujitsu ScanSnap and the Pentax DSMobile 600. The ScanSnap series don’t use TWAIN drivers so they can’t blow up ReceiptWallet and the Pentax DSMobile has incredibily well put together drivers. My guess is that they didn’t start from legacy code....

Things are so ugly out there I think Apple needs to start certifying device drivers. Vendors would then be incented to invest in drivers that earn certification and the right to advertise compliance.

Googaby: Novel approach to Google Contacts and OS X Address Book integration

From Macintouch, emphases mine:

MacInTouch: timely news and tips about Apple Macintosh, iTunes, iPhone and more...

... Mac-Chi released Googaby 1.0, a Google Contacts and Address Book utility for Mac OS X. The software enables you to drag-and-drop contact information from any application that allows selecting text and then parses the data, presents it for review and editing, and stores the approved version in Google Contacts. It can also push new or changed Address Book contacts, including pictures and group memberships, to Google in the background. Other features include backup and restore of Google Contacts and Groups, drag-and-drop between Google accounts, drag-and-drop into spreadsheets, support for executing custom AppleScripts, a hyperlinked Note field, and more. Googaby is $24.95 for Mac OS X 10.4 and up (Universal Binary)."

I don't know of any other solution that can backup or interact with Google Contact Groups. They have a somewhat limited 10 day trial solution -- a bit short for my tastes.

Note, however, that Googaby doesn't attempt to synchronize Google Contacts with Address Book ...

... In the some of the articles which have been published so far Googaby and “sync” have been mentioned in the same sentence...

So I want to be real clear: Googaby is NOT a sync agent nor sync tool!

Googaby PUSHES changes from the Address Book to Google Contacts. Pushes one way and only one way.

If you’re a Mac user then you should store your contact data in Address Book.
Not only are the data elements in Address Book much richer than those of Google Contacts, but by storing your contacts in the Address Book...

I'd love true sync between Address Book and Google Contacts, but I think the data models are too different to really do that. This seems plausible.

I'll give it a cautious try on a secondary Google account where I can do some testing safely. (It's easy to backup and restore the OS X Address Book, there's NO backup for the Gmail Contact/Group collection.)

Update: I've found my first (minor) bug. If you launch Googaby from the first time and don't have a network connection, it hangs instead of displaying an error message or log-in dialog. It should display an error message. I had to kill the process.

Update: I'm missing something. It can see how it does a backup and restore of Google Contacts, though I've not tested that. I can also see how it might work to better manage one's Google Contacts, and have them available offline -- but since Googleby is the "source of truth" then any Google Contact changes made, say, by Google Sync to the iPhone will get lost. I also don't see how one creates a relationship between an Address Book entry and a Googleby entry. I see a lot of complexity in the Preferences settings. There's nothing much in the Help files. Googleby looks interesting, but it shouldn't have been launched without a manual of some kind.

Aperture keyword hierarchy does not support inheritance queries

I use Aperture in a very odd way. I use it to process my Canon RAW files before I convert them to JPEG and store them in iPhoto.

I then delete the originals.

Weird, eh?

Problem is RAW files cannot be used to archive images, and I really care about longevity. I'd prefer a better file format, but for I've been thinking JPEG is the best we have*.

In addition, when I bought Aperture I thought I'd migrate from iPhoto. Alas, Aperture does not support metadata associated with collections of images (albums) -- and that's important for us. Importing of iPhoto Libraries into Aperture discards important information. So I've stayed, grumpily, with iPhoto. (My Apple love died some years ago - no surprise to readers of this blog.)

Using Aperture as a kind of darkroom means I've never done much with Aperture keywords, but on a recent solo non-work related plane flight I took my Aperture Library along to catch up on image processing. I had a rare couple of hours to look at Aperture's keywords.

The first thing I noticed is that I could organize them in a tree (hieararchical, outline) structure using the (shift-H) Keyword HUD. It's a strict tree-type directed graph -- no mulitple inheritance. Any term (node) can be a keyword (the user interface and documentation are misleading on this).

Wonderful -- I thought. I'd prefer an acyclic directed graph with multiple inheritance, but a tree is nice. I can create an ontology of family relationships and execute inheritance queries -- so a query on "FL" will find all children of "FL" such as Emily, Ben, Brinna, Tim, Kateva, me, etc.

Wrong.

The tree is merely a way to organize terms. As far as Aperture is concerned the tree nodes (terms) are flat keywords assigned to images. All queries are term queries -- no inheritance, no subsumption.

Blech.

Oh well, it does make it easier to continue my disgruntled data locked iPhoto relationship.

* I would love it if Microsoft's HD Photo (superb LOC description) were to be blessed by the Library of Congress. The LOC currently favors JPEG 2000 for lossy compressed archival images (NDNP profile specifically), but their next best choise is JPEG - which is what I use. I've another post due on this topic. After reading the LOC recommenations I need to reconsider whether JPEG is still my only effective archival option. (Update: Aperture doesn't export JPEG 2000! Weird, since iPhoto can manage JPEG 2000 and so can Quicktime. Looks like I'm staying with JPEG for now ...)