My father still uses dial-up. I signed him up for an AllVantage dial-up account because it was only $5 a month. Recently AllVantage got bought by PeoplePC. There were two problems. PeoplePC charges $11 a month, and they require you to use their Windows only software for connecting. It seemed like a no go, so I called to cancel. The woman offered to only charge $4.47 for six months, I said no thanks, so when the woman said, $4.47 a month indefinitely, I figured it was worth figuring out a way around the requirement for the Windows software.

I first emailed their tech support address. I asked if I could just use PPP on a Mac. This is the response I got:

“Regarding your concern, we suggest you to reload PeoplePal Toolbar.”

I wrote back saying I just wanted to know about PPP and Mac support. I got this back:

Regarding your concern, we would suggest you to first connect to PeoplePC Online service using the PeoplePC dialer and then try accessing Internet explorer and Outlook Express.

That route was clearly a dead-end. I decided to just try it. First, I tracked down a local access number. You get a list here. (I had erroneously been looking for them in the support documentation.)

The rest turns out to be remarkably simple. Your username is simply your PeoplePC email address (username@peoplepc.com) and your account password.

With these three items in hand, PPP dial-up works flawlessly.

Most people know that Fiber Optic is has much higher capacity than copper lines, but I was really amazed from the comparison made in the “Light Speed” episode of the PBS show “Innovation”.

That red line on the right of this photo is a single strand of fiber optic cable…

It has the same bandwidth capacity as all these bundles of copper:

Over the years I’ve written dozens of perl scripts to scrape HTML pages and generate RSS feeds. Then I tried RSSxl which was pretty kludgy and lacked features. Recently, a few others have come out including FeedYes and PonyFish.
But none worked very well. They weren’t flexible enough, or the interface was poor.

Feed43 is the first to impress me. It is fairly powerful allowing you to match any number of different elements on a page and then create a custom template for output. The interface is also excellent. You do need to read the documentation, but after about 10 minutes I was creating RSS feeds to my exact needs.

The Feed43 developer is committed to keeping the service free and make money by having a premium service with extra features. The only down sides that I’ve run into so far is that the feed only gets updated every 6 hours which his fine for pages that rarely change, but much too infrequent for pages that get updated constantly throughout the day. The regular expression matching is also a bit limiting. Instead of full regex support, it has reduced matching to a handful of tags. So far, I’ve been able to make it work, but on nastier pages that need complex matching, I suspect youd’ run into a dead end. For instance, I don’t how it deals with greedy matching.

Here’s a few screenshots of the process to make a feed of the Brattleboro Reformer front page news headlines:

Step One: Enter a URL

Step Two: Write the (ir)regular expression to match your elements (title, link, description)

Step Three: Extract your elements and verify you’re getting what you expected.

Step Four: Make an output template. This one is quite simple, but you can get as fancy as you wish… putting multiple matched elements in the title, link or description.

Step Five: Preview your handywork. Optionally give the feed a pretty name, password protect it, etc.

Step Six: Subscribe in your feed reader of choice.

Follow up to my post about Flickr and adding the original file name in directly in the Exif data… After some brief skimming of Exif documentation, I didn’t see anything that suggested I could add arbitrary tags.  I tried doing it with the exiftool utility and it doesn’t allow it.  So instead I added a “comment” tag with the contents: IMG_1234.jpg and that worked.  I then uploaded the image to Flickr and viewed the Exif data both through the Flickr website and through the Flickr API.  Oddly, the Exif comment data doesn’t appear on either the Flickr website, or through the Flickr API.  It does however preserve the Exif data, so if you download the original image it will still be there.

The solution of embedding the original filename in the Exif data won’t work anyway since I can’t easily add the Exif tag to the photos I’ve already uploaded to Flickr.  So it looks like a regular Flickr tag will be the best solution for now.  While it is a little verbose, I think I’ll tag them with the above formatting for easier parsing and clarity.

By the way, the Flickr Creative Commons page was finally fixed a couple days ago.

I’ve decided to slowly migrate the redjar gallery from Menalto Gallery to Flickr. Gallery just isn’t doing it for me. Feature bloat in the wrong areas and regular security vulnerabilities top my list. Flickr has it’s own downsides, loss of control and flexibility, but with a little work I think I can work around them. The benefits of Flickr are that I don’t need to worry about maintenance. They get to worry about things like security vulnerabilties and backups. And of course there is the community. For instance, if you are looking for a creative commons licensed photo of Amsterdam, or puppies, or whatever, Flickr is a great place to start.

It is going to be a mult-phase process. So far, in the past couple days I’ve migrated over 6500 photos in over 300 albums from gallery to Flickr using the gallery2flickr script. It fails to export the album title and description, so I had to take care of that with my own combination of scripts, Flickr API, and some kludgy TextWrangler find and replace.

One thing I haven’t taken care of yet is addressing the fact that when you upload an image to flickr, the original file name gets blown away. I can understand them changing the name so that all photos follow a convention, while avoiding name collisions, but there really should be a metadata field accessible via the public API that stores the original filename. My planned solution is to add a tag for each photo in the form of, “original_filename_IMG_123.jpg”, where IMG_123.jpg is the original filename. Another alternative that I haven’t looked at is the possiblity of stuffing the original filename in the EXIF data and seeing if it is maintained. I’ve never played enough the EXIF to know how straight forward that is.

The next phase, and the key to my even considering using Flickr is to use the Flickr API to put the photos on redjar.org, hopefully without the viewer even really knowing that they are hosted by Flickr (beyond the API requirement that there be a link back to the photo on Flickr).

I’ve looked at a couple projects that do just this including f*gallery (no CMS dependencies but also not very advanced and not being actively developed), FAlbum (Wordpress plugin) and this Flickr module for “CMS Made Simple” which is the closest to what I’m after, but of course is tied to “CMS Made Simple”. However, none of them go quite far enough. I don’t even want the image src to use flickr.com, I still want it to be a redjar.org URL. Why? Because I am absolutely sure, that in a few years, the next Flickr will come along that will make Flickr’s offering look rediculous, and I’ll want to switch, hopefully without breaking URL’s.

By the way… strangely, after my Flickr account reached 100% of the 2 GB per month upload limit, my uploads just kept chugging along. Weird, but I’m not complaining.

What is incredible about Google is that they can initiate a project and make it public, before I can even make a blog post about a vague idea I have. The following blog post has been in my unpublished blogs bin for weeks now. Today, Google opened Google Base. It’s been talked about for a week or two now, but it is actually closer to what I’ve been dreaming about than the rumors led me to believe. It still has a long way to go. Most importantly the need for a public API… which I’m sure will be available soon. I most excited that Google Base will spur other similar projects. On with the dusty old blog post:

For many months now, maybe I even have a blog post about it, I’ve been wanting a web application that would basically be Filemaker for the web, combined with Wiki like collaboration and versioning. Everyone I’ve tried to explain this to looks at me funny. Either because it is just a stupid idea, or because people don’t see the brilliance of the concept. I think I’ve finally convinced Seth it is a good idea, but he has his own brilliant ideas to spend his non-existent free time on.

Somewhat encouraging is that there seem to be a couple projects circling the idea, but don’t quite get what I’m after.

What am I after? How about an example. Imagine you want to build a collaborative dictionary where anyone could add new entries, or edit existing entries. You could take the Wiktionary approach by using pure wiki software. This has the benefit of ease of collaboration and every edit is saved so you can revert vandalism etc. However, what if you wanted to have a clean xml version of the data so that you could use the dictionary from other outside applications? Each entry is just a big blog. Pulling out the distinct elements (word, definition, pronunciation, etc.) is not clean.

The alternative implementation is an Urban Dictionary route. Building a custom application from scratch. The downside to this approach is that there is a bunch of coding infrastructure that must be done. In addition, making the app have flexible access control and extensive versioning make the coding much more complicated.

My proposed solution is a web app that allows a user, through a web browser, to create a new database, define columns, types etc. (basically PHPmyAdmin). Then define access control roles (ex. anyone can add a new entry, or edit old entries) with all revisions saved. Finally they can define views/layouts. So they can provide a HTML template and a stripped down XML view.

Other possible uses?:

• Television and Radio schedules
• An open collaborative Internet Movie Database or CD database.
• Database of geo-locations
• Events calendar
• A plain old Wiki
• To-do lists

No programming required. Obviously it is a very complicated application, but all the pieces exist on their own. Just need to put the following projects in a blender:

• WikiDB
• MediaWiki
• phpMyAdmin
• base.google.com
• WikiCalc
• Ning

I made 30 cents on Amazon’s Mechanical Turk before the site got Slashdotted and all the simple tasks got used up. I was identifying photos of shop fronts in Boston. It gives you a page with a business name, an address, and 5 pictures. Your task is to select the photo that matches the business address. In only one case was the actual business name present. All the others either were businesses in a big building with no store front, or there was just no photo that matched at all. This is just one example of uses for Mechanical Turk. It is an interesting idea… one that we were actually discussing after the Berkman luncheon the other day. It could actually be used to build a seed database for Artificial intelligence systems to use, which Mako says MIT is working on.

One interesting thing that I came across in the photos themselves is the vehicle that was used to take the store front photos (see photo). It appears to be nothing more than an SUV with a video camera on top. The photos are used in Amazon’s A9 map. I find them useless in most cases. The photos are head-on and very tight which gives you a very limited view of the building. They really should taking the photo at an angle and with a wider angle lens. All that trouble they’ve gone through, you’d think they’d try to do it right.

A bunch of us PRX’ers along with some old fellow Hampsters (Ben Mauer, Mako Hill, Erik Hopp, and Kellan Elliott-McCrea) went to the Berkman luncheon on Tuesday. The guest was del.ico.us founder Joshua Schachter.

Joshua started his, “presentation” by, “well, I didn’t prepare anything” which is never a good sign. From there is was pretty much all question and answer which is good, except the questions from such gatherings are often a mixed bag.

David Weinberger has a write up about the session.

I’ve been keeping an eye the Internet Archive’s recent project to digitize books in collaboration with several other companies and academic institutions. Two thoughts, one related, and one not really.

Why does the project not ever seem to mention Project Gutenberg and the sub project Distributed Proofreaders. I understand that the IA project is different in many ways, but it is a real bummer that PG doesn’t get props, and isn’t involved in the project. I would imagine they could provide some valuable insight, and with Distributed Proofreaders, some valuable resources.

My other thought is one that has been rattling around in my head for a while. Internet Archive is a incredibly cool project. Kahle’s vision is to build a modern day Library of Alexandria to preserve important cultural knowledge. In his presentations he always mentions the need to store the data in multiple locations around the world to avoid the library of Alexandria’s fate (destroyed by means that are not quite clear). It turns out that culturally significant works are already being preserved, and in not just one place, but 10’s of thousands of copies all over the globe. Is it librarians doing this work? Nope, it is the “pirates”. If any work has significant cultural value, pirates will make copies. In a x hundred years when the IA is a faint memory, researchers will come across some pirates ancient hard drive full of DivX movies. And of course, the pirates have a huge advantage over Libraries, Internet Archive, Google print, etc. They don’t have to go through the politics and details of copyright and the DMCA.

I’ve used populicio.us for a while. It provides an RSS feed of recent popular links posted on del.icio.us. However, most entries are just the a title and a link. Without actually clicking the link you can’t even tell what the link is about. And without any context, I’m not very likely to click the link.

To add context, it would be cool to use the link provided by populicio.us, grab the linked page and include the actual page content in a derivative RSS feed.

I’m one step closer to cheap Internet connectivity on the commuter train. Today I picked up a USB data cable from Radio Shack to connect my Samsung A670 phone to my Powerbook. After some fiddling today, I was able to get a decent Internet connection using my mobile phone. I was getting a respectable real world download speed around 10-15 Kilobytes/second. (bytes not bits)

After a lot of conflicting opinions, I’m fairly certain that this service only uses my allocated monthly minutes. When I logged into Verizon Wireless, my account says that “National Access” is enabled. I didn’t pay anything extra for it.

There are many good instructions on the Web, I used a recent how-to from Engadet. But here is my basic set-up (running Mac OS X 10.4 Tiger) with Verizon Wireless, a Samsung A670 phone in the greater Boston area. My phone has a “1X” icon in the upper right which I think is important for connecting at the higher speed.

• get USB Data cable ($22 from Radio Shack, I bought it locally so I could just bring it back if it didn’t work)
• connected phone to Powerbook with the cable. Initially nothing happened. After turning the phone off and back on and quiting the system preferences it appeared.
• Set the phone up as a new network device.
• Set the Modem type to “Verizon Support (PC 55220)”. First I tried the Verizon_Wireless_STD_Driver which I downloaded from the web as instructed from the directions I was following. However, the modem would connect and then immediately disconnect. The Verizon Support driver was included with OS 10.4
• Set the Account Name to: myphonenumber@vzw3g.com
• Set the password to: vzw
• Set the Telephone number to: #777
• Connect.

The obvious downside is that Corin and I only have 500 daytime minutes to share each month. No where near enough if I were to connect every day on the commuter train to and from work. However, Verizon allows unlimited calling minutes to other Verizon phones, AND they provide additional phone line for $10 a month. I’m curious if it is possible to get another mobile phone, connect it to my Internet connected server at home. Then, instead of using Verizon as the ISP, dial the other cell phone from my laptop, and connect to the Internet through my home cable connection. This would allow unlimited internet connectivity.

A couple issues have already sprung to mind. How to get the recieving phone to pick up (can the OS control this?) and the other is keeping the phone charged. I was surprised to see that the USB cable I got does not charge the phone when connected to the Powerbook.

We’ve had RCN Cable Internet and TV at our new place for about a week. The claim is 10 Mbps down, 800 Kbps up. With TV, it costs 45 dollars for the first 5 months. For most tasks, it doesn’t feel any faster than a T1.

Direcway isn’t quite out of my life forever yet. I have to cancel the service. Because their service sucks so bad, they require customers to enter a 15 month service agreement. I’ll be canceling before 15 months has elapsed which means a lovely $300 fine.

My morning commute on the train is also making Verizon’s EVDO service look very appealing. But even after they dropped the price from $80 a month to $60, it is still too price considering I’d only be using it for an hour a day. I’m keeping my eye open for people on the train that are regulars and always have their laptop out. Then thinking of proposing that we split the cost and we set up an adhoc network with the EVDO laptop sharing the network.

A couple days ago I actually sat next to a business man who pulled out a laptop, then pulled out a PCMCIA card with an antenna sticking out of it. I asked him if it was a EVDO card and he said he wasn’t sure. Apparently the IT department had just given it to him that day.

He plugged it in, connected to his office via a VPN, and seemed to have uninterrupted connectivity the whole way home.

I went to watch a screening of “Blogumentary“. A documentary about blogs by Chuck Olsen. Chuck was on hand to answer questions afterwards. The film was created entirely by Chuck over a period of two years with 10,000-20,000 of his own dollars.

Film synopsis:

Blogumentary explores the many ways blogs are affecting our media and politics. Personal political writing is the foundation of our democracy, but mass media has made us into passive consumers instead of active citizens. Blogs return us to our roots and reengage us in democracy.

The film features interviews with Jeff Jarvis, Joe Trippi, Rebecca Blood, David Weinberger, Dan Gillmor, John Hinderaker and many more. The film also explores the personal side of blogging, including a BBC blogger injured in Iraq and a friend’s suicidal blog post. Whatever your familiarity with blogs, Blogumentary offers a fresh and compelling look at how blogs are changing the way we communicate.

After hearing good buzz about Ruby on Rails for a while, I took the opportunity to learn more about it by going to the Boston Linux & Unix User Group meeting on July 20th. Rajiv Manglani was giving a presentation of his experience. He had only been using Rails for 2 weeks, but was able to give a pretty compelling presentation.

Seth has also been wanting to try rails, so we purchased Agile Web Development with Rails. Now we just need the time to read up and try it out. I got it installed on my Powerbook and started playing with it this weekend. I only managed a couple hours, but so far it looks like a refreshing way to develop web apps.

Catching up my city life… Back on July 14th I headed over to the Berkman Center to see Dave Winer show off his OPML application. (Berkman page with more info)

Two weeks later, the application has been released, but I haven’t had time to try it myself. Outlining isn’t exactly glamorous and the OPML editor isn’t very polished yet, but it looks useful for several types of applications.

Odeo (Blogger creator Evan Williams new podcasting startup) has opened its doors to the general web public. It is overall pretty slick. Unfortunately the create section is still not ready. This part distinguishes Odeo from others, so hopefully they make it live soon. Also interesting is that my Unofficial StoryCorps Podcast is a featured channel in every category that it is tagged under (ie. Documentary, Public Radio, interviews)

For every Microsoft announcement of a product/service that will be available some months/years down the road, Google seems to announce a product/service that is available immediately.

Today they announced an official API for Google Maps. This is finally a reason for me to get off my butt and learn the ins and outs of Javascript and XML.

After work today, Seth and I started playing around with the API. Recently Seth worked on a greasemonkey script to show MBTA subway and bus lines through google maps. This time around we decided to try overlaying the T lines on top of a regular map using the API’s built-in polyline functionality. The benefit is that you get to use the excellent street and satellite data from google, and no greasemonkey script is required. Seth started putting some code together while I figured out the latitude and logitude of the various T stops in Boston. So far I only have a few of the locations of the red line, but it quickly became clear that it would work.

Click here for a working example. You can click each T stop to get it’s name, and in the future, any other relevant information. I’ll keep gathering T stop locations so we can add other T lines. Adding bus lines might get a little crazy.

Now that the API is official, I expect to see some pretty amazing applications built on top of it. All the while Microsoft is working on getting RSS into IE/Longhorn which should be out any year now.

RSSxl is a cool way to make a quick RSS feed from a web page that doesn’t provide its own.

I’ve set up a couple feeds and using it for a couple weeks with excellent results. For example, the local news paper has a web site but no RSS feed. It looks like RSSxl also supports regular expressions, but I’m not sure it it allows you to store the expressions for later use. For simple pages, it sure beats writing a page scraper.

Flickr has come up in several of my recent posts, so it probably isn’t a surprise that I’m being sucked into the warm glow. The community and numerous features are really intriguing to me. RSS feeds everywhere, extensive use of tags, public API’s, starting to use AJAX instead of Flash, Creative Commons license integration, and many more nice touches make it a great product. I was just about ready to plunk down the $25 for a one year pro account. But first I started looking at the details and have decided to continue hosting my own photos using Gallery. Perhaps OurMedia.org or Google will make things a bit more competative.

My Flickr issues:

The Terms of Use are mostly benign… for now. However, they right up front say they can change the Terms at any time, in any way, with no notification. This is pretty standard these days, but I sort of expect better of Flickr. The most annoying piece of the Terms of Use for me is:

The Flickr service makes it possible to post images hosted on Flickr to outside websites. This use is accepted (and even encouraged!). However, pages on other websites which display images hosted on flickr.com must provide a link back to Flickr from each photo to its photo page on Flickr.

I would really like to take advantage of the public API to suck my pictures into my site so that people wouldn’t even necessarly know that they came from Flickr. I wouldn’t mind having a single link to flickr.com on my external page, but making every single photo link directly to the Flickr photo page is way over the top. I’m not sure I really understand their motivitation for this requirement. It they want to motiviate people to come to and use Flick, why not just require a Flickr graphic that links to Flickr.com. (Sourceforge does this.)

Another downside, is that subscription services have a way of rising prices, and when you no longer feel like paying up every year, you are left with nothing. It isn’t quite as dramatic with Flickr, but things do sort of erode over time. If you stop paying for a Pro account, your photos won’t be deleted, however, your account will become a normal free account which limits your photostream to the 200 most recent photos. Within Flickr, your older photos will effectively disappear. Also, if you don’t log-in with a free account for 90 days, the account will be completely deleted.

Then of course are Ads and the role that they’ll play in the future. Flickr states that, due to “large surges in bandwidth can be very expensive”, they reserve the right to display ads to non-members when they view your photos, “in extreme cases”. Extreme cases is undefined.

None of these issues alone are show stoppers, but I had to search over several pages to find these details and together it just feels to early to move my 12,000 photos over.

A new radio program called, “Open Source“, hosted by Christopher Lydon is about to be syndicated by Public Radio International. They’ve posted pilot shows on the show website. I just finished listening to Pilot #3 about Wikipedia. Jimbo (wikipedia founder) is along for the entire show. They also talk with a couple of librarians that make we want to hit myself in the head with a sock full of nickels.

They are critical of Wikipedia because:

• Their students are going to wikipedia and using the information as holy gospel.
• Their students are going to wikipedia and plagerizing it because it doesn’t appear to have an owner.
• Every Wikipedia entry is not complete.
• Wikipedia entries may have information that is incorrect.

These same arguments come up in every Wikipedia story and it just drives me crazy. Rewind a few years and this same exact argument came up with the Internet. If you want to feel like you’re living in the dot com boom again, just replace Wikipedia with Internet when the librarians talk. I guess you could replace Wikipedia with any messenger of information.

Librarians, your argument doesn’t change anything. Wikipedia and the Internet are growing and here to stay . Why are you wasting your time complaining when you could be educating your patrons how to think critically. Teach them the difference between an encyclopedia and a primary source. Teach them to check their sources. They themselves bring this up in the interview, and then go right back to dissing the Wikipedia.

The root of their argument is of course their fear that their jobs are obsolete. Well wake the heck up, your job is obsolete. It doesn’t mean you are out of a job, it just means you need to update your job description and the role a library plays in a community. Libraries in many ways are great (I love libraries), but in many other ways they are also inconvenient. The Lisa Simpsons of the world may continue to visit, but for the Bart Simpsons of the world, the library is nerdville and might as well not even exist.

What hurts my head more than being hit with a sock full of nickels, is that I think the librarians agree with me, but because Wikipedia is not in fact the holy gospel of all human knowledge, instead of saying, “it is really cool and helpful, but should be used critically just like any other source”, they seem to suggest that it is just a toy that shouldn’t be used at all. The same could be said for books. There are really crappy books out there. Ones with factual errors, ones with information not appropriate for children, ones that don’t have sources listed. Books are a fun project for its authors and readers, but the reality is the only reality is multiple reality. The only neutral point of view is as many points of view as are available.

Update: I meant to point out that one of the comments on the radio open source website points out the excellent: Wikipedia:Researching with Wikipedia page.