Thursday Apr 02, 2009

A second look at Second Brain

I have revisited secondbrain recently and decided I need to get to grips with its libraries and collections. I am not sure of the differences and whether I should be create broad large collections such as travel or software, or even something narrower but broad such as database, or use it for more project orientated collections such as specific journies or personal engineering tasks, a bit like what it takes to justify a new snipsnap page on my bliki. I quite like the fact they give me a domain name, and that I got their early enough to get "davelevy".

It now takes a much broader range of feeds, which was the criticism I made last time I reviewed it, and creates an aggregated tag cloud. This is neat, but I hadn't realised how many tags my picture collection generates. The tag cloud is dominated by the places tag and the geographic qualifiers. When you add the bookmarks created while planning the travel, it dominates the tag cloud, which I am not sure is what I want. (I wonder if they could or should permit us to weight the tags by feed.) Usability is also inhibited here because like most people, I don't tag a feed as belonging to itself, so my bookmarks aren't tagged as bookmarks. Also several of my feeds are not tagged at all. All-in-all, this is a feature I like, so I'd really like a tag cloud widget.

I was looking at second brain to see if I could make it my home page and consolidate the various sites I am using into one place, it could well be possible. I'd loose control of my look and feel, and I'd need to consider how to host original textual content but a blog might work for that if I have SB's collections and libraries. Perhaps I'll try and migrate one of my travel pages to SB and see what it looks like, and how useful I find it.

tags: ""

Wednesday Apr 23, 2008


In the article My Twitter Manifesto, I stated that I did not propose to use Twitter as a microblog, the closest thing I have to one is my feed, which since I bookmark nearly everything I read, gives you an idea of what I am thinking about at least. I nearly always comment on a site/page, although not always in my own words. But if your interested.... it's there. The most recent seven items are also in the sidebar on this page.


Friday Aug 10, 2007


Dave Levy's Facebook profile Friends are moving to Facebook, so yesterday I had my first look at it. I checked out some of the add-on applications available by browsing the utility applications pages and discovered some amazing facebook applications. I have installed "my RSS" and connected to my bliki's rss stream (No I havn't there seems to be a quality issue with it, fortunately, there are several other apps that claim to do the job). I have also installed the (which works) and the plazes plug in (which has some problems), but there are so many that seem worthwhile and I need to go back to check them out.

I should probably get some feed back from my correspondents as to how they feel about the traffic created by these applications which all seem to write into the facebook feeds.

I have connected this blog to my profile and this may mean that I need to using the summary/content feature of roller which until now I only use occasionally.

Whoever argued that opening up the applications interface is awesome competitive advantage is absolutely correct, a true proof point that free and open source creates value.

I have bookmarked "Define Me" & "Books" (now replaced by a query on books), but also think that "My wish lists" (amazon wish lists), Friend Cloud, Memberships all seem worth exploring.

Talking to Alec Muffet, he suggests that it's possible disadvantage is that it is an internet end-point, it only broadcasts to your friends in the community and none of this interesting stuff about oneself is available on the net. I suppose your friends havn't given permission to publish their content. Alec, however, will probably let us know what he thinks soon.


[Read More]

Saturday Jun 23, 2007

Plazes 2.0

Plazes, which I wrote about here...  just updated, actually a couple of days ago,  their agent and presumably the back-end infrastructure and proved the Web 2.0 permenant Beta adage as the true as ever. Its an interesting problem with their architecture that they need to update the laptop clients. I hope they have improved their network update functionality as the automatic update failed to work on my XP client and the manual downloads did not become available for a couple of days. I was still able to use the phone to update the web site though so no catastrophe.

The web site functionality seems much richer so thats good.

To me one of the biggest outstanding problems is that they don't stimualte a self-help community. It would be easy enough to copy the games companies, or even Dell and Sun's developer forums. Its sort of interesting who does and who doesn't. For instance I can't find a Sony-Ericcsson official community site but have found an unofficial bulletin board for Sony Errison users. Flickr has a help site although I have only found two forums, New Ideas and everything else. My favourite of the official sites are the games sites of which Bioware Forums are a good example offering personalisation tools to allow people to track their own interests. Several of the Bioware fan forums use Invision's community management software. Two implementations that illustrate the strength of community and the personalisation capability within this tool  are Gibberlings 3 and Spellhold Studios, while RPG Dungeon uses Simple Machine Framework. As ever, you need to read the small print to discover which are free and open source. Both these solutions have RSS.


Tuesday May 08, 2007

Where am I now?

During my last trip to the USA, I took out a subscription to plazes, this'll allow people to know roughly where I am. They have a flash badge which I have inserted onto the sidebar on my "About me!" page.

Its also here...

Download Flash plugin

Unfortunately, I can't find the Linux version of the plazer (their client agent), despite having asked them for a pointer, and attempted to 'google' it  and so given I'm mainly using Solaris as my desktop, which they rudly describe as 'legacy' the timeliness of the geographic data is less than it might be, but it should give you an idea of where I am. Let's face it, a bit of inaccuracy is probably good, close enough for my friends, far enough away for the rest of you :)

They distribute agents for windows and the mac, and they claim Linux. I propose to see how good the Solaris Express Linux compatability offerings are when I get myself a Linux agent. 


Thursday Nov 30, 2006

Tags in Roller & Technorati

Two days ago, I looked at how to display the tags assigned to an article hosted in roller and left the question of generating techorati tags unfinished.

<!-- Now lets technorati tag the tags -->
<P><small>technorati tags:
    #set ($mytags=$entry.tags)
    #foreach ( $tag in $mytags )
        <a rel="tag" href="$">$</a>
        <!-- NB there is a trailing space on the line above #end -->

should produce the technorati tag list in a small font paragraph. Now on to Chris's challange.  

I have a challenge about how to migrate forward, since if I put this code in the _day file (inside the entries loop), than I will get the technorati tags twice, for those articles with both roller tags and technorati tags specified. This is caused by my eagerness to load my tags into roller to see my tag cloud; thats why I have articles with both roller and techorati tags. I might be able to create a local velocity macro and use this as content within an article, but I think the _day file is the right place. Have to think about it.

The tags are thus still done by hand.



Monday Nov 27, 2006

Tuning tags for query & propagation

I am still hand coding the the tags in this blog. Roller has a macro to display a hyperlinked tag list but resolves the links inside the roller instance.

<P><small>tags: #showEntryTags($entry)</small></P>

However, I still want to use technorati, so I am copying the tags from the input text box into the content body and using the topic plugin to create my technorati tags. (Actually I don't even necessarily always keep the same tags, but I am pretty consistent.) I wonder how to write velocity code to build the technorati tag list.


Friday Nov 10, 2006

Tag Clouds & Tagrolls

I have a new feature on the Blog called More Tags & Links, currently labelled as Beta. 

This has my roller tag cloud and also a tag roll from I hope you find it interesting.

Due to screen real estate problems I have removed the page menu from the main panel on the blog front page. It remains at the top of my front page sidebar.


Thursday Nov 02, 2006

More about Roller tagging

I have just been experimenting with the tag macros. It is possible to display an articles roller tags using the #showEntryTags($entry) macro. This can be inserted in the entries loop in _day file. 

Because I am planning to continue to use the technorati tags, unless I plan to use a hidden display format I would have two copies of the tag list in the HTML rendered version. So I have not implemented this macro in the _day file. 

I am struggling to get the tag cloud displayed, it seems to refuse to interpret the macro call

The #showTagCloud macro was not implemented since the velociraptor code is so simple (yeah right, I couldn't work it out). I now have a tag cloud implemented on the "About Me" page; Dave Johnson gave me some sample code to make it work. There were other reasons for not implementing showTagCloud as a macro. The team were unclear given the number of customisations that would be required, how used the macro would be, because no matter how few used it, it would need to be maintained.

First set up some CSS rules, roller offers five levels of weight, so five rules is good. 

          .s1 {font-size:60%;}
          .s2 {font-size:80%;}
          .s3 {font-size:100%;}
          .s4 {font-size:120%;}
          .s5 {font-size:140%;}

These are relative rules, which is jolly good so they inherit from the Body or Paragraph rules, but you can make the rules as complex as you want. The following velociraptor/roller code implements the tag cloud in the html file.

          #set($mytags = $model.weblog.getPopularTags(-1, 100))
          #foreach ($tag in $mytags)
              <a class="tag s${tag.intensity}" href="$url.tag($"

Note the way that the CSS rule names agree with the class attribute of the anchor tag.

This article was updated at 13:30 GMT 3rd November GMT.


Tagging comes to Roller

Yeserday,'s "Roller" implementation was updated to V3.1. The roller projects page is here.... This has integrated tagging and the Xihna Editor. I had to change the editor in my settings page and this is the first article written using it.

I wonder how to apply classes to the paragraphs, so

dave B$ whence html

was tagged by using the show html source button and inserting the class attribute by hand.

The article is also the first article with roller integrated tagging, so I'm not yet sure about how to ensure that technorati tags and roller tags are the same and the tags are posted to the relevant sites. So we may have two sets of tags on this article.

The technorati tags are


and I have repeated them in the roller entry tag edit box.

I may go back through the blog and tag every article, or at least all those I have already tagged, and I need to work out how to display my blog's tag cloud


Thursday Aug 24, 2006

Using to find Flickr pictures

Did you know that if you bookmark a flickr picture at, that it posts a thumbnail into the HTML page of the feed.


screenshot graphic


I originally wrote, if you check my delicious HTML page out now, you can see what I mean, but I decided to upload the screenshot. I have however checked the RSS feed using Mozilla, RSS Owl and Bloglines, and it would seem that the picture is not propagated inside the RSS feed.


Tuesday May 23, 2006

Tags and Spontaneity

We broke for coffee and I bumped into Elias Torres, one of the roller development team, so we had a long chat about tags on the web , looked at how to make roller's implementation better and how I could become more helpful by creating an environment to test this stuff. He pointed out that my Qube (currently not working) needs replacing.

We tried to remember the state of our debate as we left it. Elias co-authored a paper "Ranking Bookmarks & Bistros", one of the conclusions is that tagging software should not be too helpful in prompting for tags; the spontaneity of the tag author is an important part of the folksonomy. My view is that a tagging correspondent will have a personal structure to the tags they use, in my case I have a small group of 1st order tags, which will always be used, or nearly always. I believe that my contribution to the folksonomy is more effective if I am helped to keep to my structure. For my medium order tags, I want to be prompted as to my spelling and which words I use to imply meaning. Elias' argument is that the collective expression is stronger if more spontaneous (I think). I think that part of our disagreement is that I focus on tagging for personal use and at time of discovery and hence initial tagging, Elias is concerned about the stranger looking for wisdom in the tag base.

For instance for my retrievals I like to go in large or medium and refine my queries using the "+" function adding filters to my query until the page is small enough to inspect. (Is this a binary chop?) It all reminds me that I need to ask him for the data model (author, articles, articles, tags) numbers and growth rates. I wonder if Linda or Will can give them to me from





« August 2016