This blog is moving back to tumblr

I tried, I really tried, but, I just can’t get into blogging using WordPress. I can’t explain why. I think part of it is that while there’s many, many features you can customize, you can’t (easily) get into the guts of the templates. Furthermore I like to play with things like ads and other widgets, not really to make money (I’ve earned a whopping $95 in five years of advertising across multiple sites!) but to experiment and learn about them so I have a slight clue when people ask me about them.

So, I will be shifting the site back over to tumblr.com. If I do it right then http://epcostello.com/ will still display a site (though possibly not all the content that was here) and the feed will be http://epcostello.com/feed.

Note that my personal blog/site is still at epcostello.net. epcostello.com was a notion I had to blog more professionally, whatever that means.

Zooming along at 80mph on the Lake Shore Ltd and I decide to write a twitter app

I’ve been toying with geo data for years now but nothing has ever quite clicked. Some years ago I bought a bluetooth GPS receiver for use with a Nokia phone which allegedly supported it, however none of the extant software I could find actually seemed to do anything useful with the data, keeping it locked up behind their online service.

What I wanted all along was a simple, easy to use tool to track my where abouts. I figure if the government and corporate America can track me, why not myself?

Google Latitude seemed to be a good start for this, but because of Google’s well–understood paranoia about tracking people the Latitude service is very stripped down, keeping only your current location and providing no history. The s60 app for Google Maps will update your location in the background using both GPS as well as triangulation (wifi? cell tower?). It “just works” — one of my major criteria.

So some months ago I started polling my own Latitude feed and stashing it in a directory on my server. I’ve been able to use that to craft some cheesy breadcrumb maps of where I’ve traveled since July, but otherwise haven’t done anything with the data.

I’d been meaning to set up something to update my twitter location field based on one of these feeds but have been preoccupied with other things.

Finally in a fit of pique tonight I cobbled something together on my Mac, it seems to work (within the limitations of Google’s Latitude data, which seems less precise than what you find in Google Maps, and twitter’s location field which is limited to 30 characters).

One minor thing I ran into involved my latitude id, which is a large, 19 digit number.

On my Mac, running Snow Leopard, I assigned it as a number and didn’t really think about it. Everything worked fine.

I uploaded the script to my hosting provider and consistently got bad data out of Google and couldn’t figure out why.

Well, my hosting provider runs a 32 bit system. The numeric id was overflowing and creating a negative number for the userid, so Latitude returned a location of 0.000,0.000.

Changing it to a string fixed the problem.

So as I train westward for Defrag 2009 I’m going to let this script run a bit and see what else I can do with the data.

More observations on twitter lists

This is a running followup to Thoughts and comments on twitter’s new lists feature. I’ll keep updating this post rather than create new posts.

• You can add your own account to a list you’ve created through the API, but not through the web UI. Once added, you cannot remove your account from the list through the web UI, only through the API. (19/10/09)
• Lists do not show any replies, even from yourself. For example, I have a list called @epc/defrag which includes my id (@epc). The list does not show my own replies to others, though the replies appear in my own timeline. What I can’t tell is if lists would show replies between members of a given list.

Thoughts and comments on twitter’s new lists feature

I think twitter’s lists feature could be a major addition to the service. One of the interesting creations of twitter has been asymmetric follows: there is no burden or commitment to follow back people who follow a given id. The lists feature takes this a step further: I can create and curate a list of IDs I’m interested in, and rad their tweets, without following the ids in twitter jargon.
The followed accounts will be able to see that they are in a given list, but they’re not being followed.

I think we’re going to need some new terminology.

There’s two immediate consequences and/or benefits of this change:

• Today if I follow another twitter account, that account can send me a direct message (DM) but I cannot DM that account unless it has followed me (a reciprocal follow). But if I add that twitter id to a list, that id cannot DM me back (as far as I can tell in the still–developing API). So this is another form of asymmetric following.
• It makes it much easier to follow a lot of accounts, if the UIs we use allow separation & aggregation of lists.

What I mean is that today I maintain two twitter accounts: one follows people, almost entirely people I’ve met or have a 1/2/3rd degree relationship with (it may be an extremely tenuous relationship). The other id follows almost entirely services plus some A–list twitter users. I embarked on this path because I was losing too many tweets from people I was interested in hearing from amongst all of the other tweets from the various services.

I didn’t want to shut out the services entirely, but if you follow five services which tweet 20—30 times per day, that’s potentially 100–150 tweets per day which can easily overwhelm the tweets you get from someone who posts only a couple of times per day.
Now I can segregate those service based tweets off into a list (and furthermore, by “unfollowing” many of them remove the hole into my attention that DMs provide).

The lists feature is not yet fully baked, but when released I think could be as revolutionary as geotagging tweets will be (if/when geotagging is launched).

Following are comments I posted to the twitter developer’s list yesterday:

When adding a user to a list, only the numeric id appears to work, eg:

   curl -k --user epc:pass -d id=7758742 --url https://twitter.com/epc/defrag/members.xml

works, but:

   curl -k --user epc:pass -d id=enorlin --url https://twitter.com/epc/defrag/members.xml

results in a 500 Internal Server error. I tried swapping “userid” and
“screen_name” for “id” and get a 400 Bad Request back:

   curl -k --user epc:pass -d user_id=enorlin --url https://twitter.com/epc/defrag/members.xml
   curl -k --user epc:pass -d screen_name=enorlin --url https://twitter.com/epc/defrag/members.xml

returns:

<error>You must specify a member</error>

(which is what I’d expect for “id=enorlin” instead of the 500).

Also, if you pass in the authenticating user’s numeric ID you get back
either a 503 or 502 (bad gateway) error, eg:

   curl -k --user epc:pass -d id=420363 --url https://twitter.com/epc/defrag/members.xml

however it does appear to add the user to the list. Note that once added, the user can’t be removed from the list via the current lists UI.

If you try to add a user who has blocked the authenticating user, you get a 403 “You aren’t allowed to add members to this list” which is slightly misleading, you’re not allowed to add THIS user to the list.

If user “A” has added user “B” to a list, and user “B” blocks “A”, that appears to remove “B” from A’s list permanently, even if the block is removed. Is that working as designed?

What are the rules for converting a list’s name into the slug/URI?
Just a conversion to “safe” URL characters?

Couple of ideas:

• allow mass additions of userids/screen names to lists, or maybe
  10-20 at a time
• allow lists to have a tweet-like tag line (the intent being to describe the list)
• don’t groan, but allow geotagging of lists. Ok, groan.

LotusLive iNotes

On a quick run through my Google Reader feeds I came across this post by Om Malik: Look Who’s Launching an Email Service which tells a tale of IBM’s Lotus brand launch of a hosted email solution called LotusLive iNotes.

So, some quick observations on LotusLive iNotes:

• The iNotes name dates back to the first Lotus Notes hack for the web, interNotes which became iNotes around mid–late 1995.
• 1Gb of storage for USD$36/year is a joke. Really, seriously guys? When Google Apps gives 25Gb for USD$50/year?
• The sales pitch seems to be positioned against Google’s GMail offering (with subtle comments about advertising in your email and data privacy), when Google Apps for your Domain is the more comparable product.
• You cannot sign up for LotusLive iNotes online, you must contact IBM Sales.
• When you try to contact IBM Sales by clicking the Request a quote for LotusLive Notes link, you get message code 40, This service is temporarily unavailable. Please try again later. with the very helpful <title>: IBM 2009/10/01 20:57:10.
• Poking around the lotuslive.com domain is interesting, it appears to be hosted not at IBM itself, but through a company called Virtela which advises us that Virtela is the Premier Global Network Integrator Enabling Business Solutions for Multinational Organizations. Now, that does not mean that LotusLive itself is hosted at Virtela, but it is an odd choice given that part of the sales pitch appears to be that the hosted service is provided by IBM.

Now, I have no inside insight at IBM, having left nearly eight years ago. But I find it bizarre that the company would launch yet another hosted offering built around Lotus Notes (there were a couple floating around in the 1990s, and Lotus itself had had a hosted option when IBM bought the company in 1995). If the offering is indeed hosted by a third party, then this is doubly bizarre.

The end, for now, with my posterous experiment

After playing for a bit with posterous I’ve decided to move http://epcostello.com/ over to WordPress.com for awhile. Posterous, for me, is a bit too simplistic. I know that’s a selling point, and I’m not knocking it, it’s just not what I want for a site right now.

I did like that I could add my Google Analytics id, but otherwise the inability to change the layout of the site at all, to add ads for example, or a twitter feed, or, well, anything pushed me away.

WordPress.com has more flexibility (though it too does not allow ads as far as I can tell, nor the Google Analytics thing, though I have not dived too far into it).

Note that I’ve been writing personal journals on web sites since approximately 1995 (though I would never claim to have been “blogging” since then, only since 2001 or so). The one frustration I’ve had over the years is while running my own site (see http://epcostello.net) gives the most flexibility, it’s also a pain especially with the advent of the Sino-Russo comment spam freaks.

So, I’ll play with wordpress for a bit and see what happens (note: wordpress.com, I am all too well beware of running my own wordpress installation).

Sanity’s Insane View of the Link Economy and the Long Tail

Years, and I mean, years ago, I signed up with the HMV Australia web site as an affiliate.  I’d found HMV to be a great source of music, and given the (then) near 50% discount on the AUD-USD exchange rate, it was cheaper to buy CDs from Australia and have them shipped to the US, than to buy the same CDs from Amazon (this was before Amazon Prime).

Now, time passed, I think I earned a whopping AUD$2.00 in affiliate earnings from HMV (I think they cut a check at AUD$100 so, I never even got paid).  HMV either shut down or shut down its website, turning over the program to Sanity.

Now, I never bothered to sign up for the Sanity.com.au affiliate site, my HMV account apparently got migrated over.  With the exception of the initial cutover, I never heard anything from Sanity about the program, earnings, links, promotions, etc.

Until this morning when I received this delightful missive:

Dear 973,

We are currently cleaning our affiliate database and have noticed that
you have not made any affiliate sales in the last 6 months. Therefore we
are no longer offering you an affiliate service and will be removing you
from our affiliate database.

As per the affiliate agreement you will immediately cease to use and
remove all links from your website to our website and will cease to use
and remove all trademarks, logos and other content or material provided
by or on behalf of us to you. Additionally, you will cease to represent
that you are in any way associated, endorsed by or affiliated with us. 

You are able to re-apply as an affiliate at any time, should you feel
that changes to your site will enhance your application.We wish you well
with your site in the future. 

Yours sincerely,

Sanity.com.au

“Dear 973” — they apparently lost my name in the exchange from HMV.

As far as I can tell, I haven’t made any affiliate sales in the last five years, let alone the last six months.

And the thing is, it doesn’t cost Sanity anything for me to provide the affiliate links. It’s what, a couple of rows in a database somewhere in Pennant Hills or wherever their datacentre is. The cost, if there is any at all, is in keeping up the relationship with me, which they’ve decided to terminate.

There’s no point in my digging through my site to figure out which links are out there (they probably link to HMV still). Sanity probably spent more time decided which affiliates to cut off than in contacting stale affiliates to push promotions, sales, etc.

When the drive starts clicking…start with the kicking

Ok, it’s not quite true but this morning when I turned on the TV to get my dose of fantasy financial advice from CNBC I was greeted by a disappointing message from our Tivo.

I didn’t take a screen shot, the gist of the message was “This tivo was configured with an external drive which is no longer present.  Presss Clear if you want to continue.”

I’d learned from a previous experience that once you press clear you’re hosed, there’s no recovery of the external drive.

So I unplugged the Tivo (which is the recommended way to reboot it).  And still no joy.  I unplugged the drive (a Seagate 1Tb FreeAgent drive) and plugged it in and heard the distinct sound of …

the click of death.

I remembered this sound from the heady days of my youth at IBM.  Certain IBM drives would get stuck, generally on reboot or when the system was powered down.  This happened once (“once”) to one of the drives backing www.ibm.com in 1995.

The recommended solution then?  Firmly drop the drive on a solid surface.

So, I thought back to that experience, unplugged everything, allowed the Seagate to spin down, picked it up gently, and then firmly slapped both sides of it (it’s a vertically mounted drive).

Plugged everything back in and …voila, the whole thing works again.

I take this as a temporary reprieve.  Tivo cautions you that any drive other than the approved Western Digital DVR extender drive has not endured the boot camp wilderness training required to support a Tivo.  This Seagate FreeAgent has been on almost constantly since 2007, almost precisely two years.  I don’t think that’s too bad, though I would have appreciated some notice from somewhere that the drive was about to go bad.

This is the third drive to fail on me in the past month.  All three were bought in the last two years and have had little or no turmoil (haven’t been moved much if at all, not dusty, with the exception of the Tivo probably only a couple of hours use per day).  I’m not sure if that means we’re cursed or if drive quality has gone down as capacity has increased or what.

And the problem with the increased sizes is the increased amount of data we end up carting around, it makes it harder to backup (I don’t back up these drives any more, I just configure larger and larger RAID arrays).

Of course, as with the last failure, this occurs the day I’m leaving on a trip.  I do plan to get a drobo once we return but that would not solve this problem with the Tivo.