Homepage
Sunday, September 28th, 2008
Sift CEO Ben Heald
So these guys over at Sift Media are heading for £3 million of advertising revenue this year. Probably worth hearing what their CEO has to say about online publishing. The first question is a bit silly:
What levels of participation inequality do you see from forum and community users? What can publishers do to tackle that more effectively?
Ben Heald takes a valiant stab at it:
You probably get about 10% of the audience that you are sending your emails to, contributing to the content in some way.
I don't think usage levels have changed that much. It matches the dynamics of a conference, where out of an audience of 100 people, you will get ten asking a questions and one or two people speaking to you after the presentation. I don't think that will change. We won't all become avid contributors to forums. Some people just want to take stuff.
His answer is basically right - participation will not be level - but at the end he goes seriously astray: "Some people just want to take stuff."
That may be true of some people but even more often, people are at different points of the cycle. There have been many communities where I have been a passive observer and then gone on to contribute substantially. Then there have been communities where I was very active and then dropped back to just a monitoring position.
If you cut off the people observing, you are eliminating a huge portion of future participants.
On the other hand Ben Heald's has some brilliant ideas on how to:
- create engaging online content in a business context
- remunerate the people writing and promoting that content
Here's what he has to say:
Writing to engage, rather than to inform, is a big switch. These worthy articles that quote all and sundry are fine and they have their place. But you don't feel like commenting on them at the end. You don't feel like engaging with it. You're informed but you feel a bit exhausted by it.
We are trying to be a ringmaster or a coordinator or a facilitator. The job of the editor is to seed audiences, stimulate and provoke on behalf of the whole site. All our editors are rewarded through the success of the sites in terms of the page impressions and users their articles generate, not how many words they write. They have a basic salary and the commission is based on the page impressions and the number of reads of their stories.
It's about engendering the right behaviours. Do you really need to write 1,000 words, or do 500 words that are more provoking? They become active on other forums, promoting their websites. You can't be successful when you're running your own walled garden.
Bingo! The problem is solved. The writer/promoter is not rewarded for a long dull article. The writer/promoter is rewarded for a successful article on objectively measured criteria.
How to transfer those terms into a smaller environment like Foliovision is a challenge, but it's certainly the right direction for online publishing.
The other issues is to hire engaging writers in the first place. If a writer has a history of penning dull work, publishable but somniforic, the odds are that writer isn't going to set the world on fire writing for you on business themes.
Still, it's a great way to force motivation for those who have the talent but perhaps not the will.
Business |
Thursday, September 4th, 2008
I was reluctant to try mobile internet. I was worried that mobile internet on my Macbook would:
- be clumsy in use
- cause unexplained crashes
- require lots of troubleshooting
- be unreliable
Well unfortunately my concerns were justified.
vodafone a1 e220 modem
Over the last few days since I picked up my trial Vodafone modem at A1 in Vienna, I've had to:
- restart my Macbook a dozen times to get the modem running again
- endure Skype flaking out on me six times while speaking with both clients and staff
- hard restart the computer after kernel panic
- uninstall and reinstall the VodafoneMCInstaller.2.09.01.00 five times
- spend hours finding and diagnosing the problem
This post is to help you avoid these issues and to enjoy troublefree use of your modem.
Point one, the best way to avoid mobile internet issues is not to use it. For most people a stable DSL or cable connection is more than enough. Fussy wireless technology is a time sink. If I could have back the hours I've wasted on Bluetooth devices, I'd have either two weeks of vacation or a couple thousand more euros in my bank account.
If you do wish to persist with mobile internet, the next best trick would be to insist on a self-sufficient modem-router. Advantages:
- DHCP routing is not going to crash your computer
- DHCP routing is not going to cause kernel panics
- you don't need to install any special - read buggy - software into your system
- If there is a problem, the mobile phone company has to troubleshoot it as any issue will be on their hardware and their network and not on your computer.
huawei e960
The big disadvantage of a mobile router, I believe you have to plug it into the wall - actually the Huawei E960 and E970 run off of USB power but do the routing via ethernet.
Alas, for the moment, the most common modem for mobile internet is the Huawei E220 HSDPA USB Modem.
If you do get stuck with one of these USB modems (they come free with your plan and lots of carriers don't offer the E960/E970 even iwith an upgrade premium - for instance in Austria, A1 doesn't offer the E960/E970 but Drei does: a very good reason to choose Drei), you will have to do some surgery on your Mac OS X system files.
Originally when I ran into problems with Skype, I thought it was bad old A1 (a cellphone company) blocking Skype just out of spite - as a threat to their business model.
It turns out the issue is much more benign. There are core issues with the Huawei Mac OX USB drivers.
What they principally do wrong is to write to the System log many times per minute, causing:
- the log to grow to be huge (instead of 50kb my system log was 153 MB at its peak, along with 23 MB of database in the asl.db file (/var/log/asl.db)
- tying up the hard drive and processor reading and writing these huge files multiple times per minute
- corrupt a kext from the E220 driver, forcing the user to uninstall and reinstall the software (three reboots! round trip)
Here is the pointless text which gets written every few seconds:
Sep 3 21:07:23 mk084020185026 kernel[0]: ::dataReadComplete: dingjianjian = 4
Sep 3 21:07:23 mk084020185026 kernel[0]: ::dataReadComplete: State:-431226754, Read len:123, return status 0.
Sep 3 21:07:23 mk084020185026 kernel[0]: ::AddtoQueue: Enter, Size = 123.
Sep 3 21:07:23 mk084020185026 kernel[0]: ::AddtoQueue: Exit, Return: 123.
Sep 3 21:07:23 mk084020185026 kernel[0]: ::RemovefromQueue: Enter, MaciSize = 1020.
Sep 3 21:07:23 mk084020185026 kernel[0]: ::RemovefromQueue: Exit, Return: 123.
Sep 3 21:07:23 mk084020185026 kernel[0]: ::RemovefromQueue: Enter, MaciSize = 1020.
Sep 3 21:07:23 mk084020185026 kernel[0]: ::RemovefromQueue: Exit, Return: 0.
Sep 3 21:07:23 mk084020185026 diskarbitrationd33: unable to probe /dev/disk1s0 (status code 0xFFFFFFFC).
Sep 3 21:07:23 mk084020185026 kernel[0]: ::enqueueData: State:-431226754, Write len:45.
Sep 3 21:07:23 mk084020185026 kernel[0]: ::AddtoQueue: Enter, Size = 45.
Sep 3 21:07:23 mk084020185026 kernel[0]: ::AddtoQueue: Exit, Return: 45.
Sep 3 21:07:23 mk084020185026 kernel[0]: ::RemovefromQueue: Enter, MaciSize = 4096.
Sep 3 21:07:23 mk084020185026 kernel[0]: ::RemovefromQueue: Exit, Return: 45.
Sep 3 21:07:23 mk084020185026 kernel[0]: ::SetUpTransmit: Tx is empty.
You can check for this on your own computer by opening up the Console (Applications/Utilities/Console) and clicking on System log.
The immediate emergency solution is to just run these two commands:
sudo rm /var/log/asl.dbsudo rm /var/log/system.log
The two system logs will be recreated immediately but will be at a manageable size again. Done periodically, you can probably keep running as an occasional user like this. (If you want to see those files to check their sizes you need to make hidden files visible with TinkerTool or use Pathfinder.)
The real solution is to stop the E220 writing to the system log in the first place.
It's easily enough done. The file you need to alter is:
/System/Library/LaunchDaemons/com.apple.syslogd.plist
You need to open it up and change this file. As neither the file or the directory are writeable, you need to use one of the following:
- terminal and vim
- BBEdit (which can write to System only directories with authorisation)
- Lingon (free utility, instructions on how to change syslog)
- Property List Editor (but to use Property List Editor you need to move com.apple.syslogd.plist to the desktop and then back again)
What you need to change are the parameters of the Syslog: what it records.
You want to change it from:
/usr/sbin/syslogd to /usr/sbin/syslogd -c
to:
/usr/sbin/syslogd to /usr/sbin/syslogd -c 0 -a
This setup will only allow emergency messages to get through. 0 is the highest of 8 levels. You can experiment with lower levels to see what the cut off point for dingjianjian messages is. If you find it, please let me know. -a just clears out the database and log file every 24 hours. You don't actually need to do that if you are at level 0 as very few if any messages will ever be logged.
The section you want to change is Program Arguments. When you are finished it should look like this:
< <key>ProgramArguments</key>
<array>
<string>/usr/sbin/syslogd</string>
<string>-c</string>
<string>0</string>
<string>-a</string>
</array><
The one major problem with this approach is that you lose your system log. If you ever do suffer system instability, the system log is a very useful tool for troubleshooting.
Something else which I did which didn't harm performance in any way is that I changed the modem from HUAWEI to Generic in the advance settings in the network panel.
generic HSDPA settings
Every time you connect or after waking from sleep, you have to start the Vodafone Mobile Connect, wait for it to find a connection, press activate:
vodafone mobile connect
Finally you connect via the menu bar modem item.
huawei e220 hsdpa connect
PERFORMANCE
When it is working, HSDPA is fast. I was getting download speeds of over 420 KB/sec and upload speeds of 40 KB/sec. That's more than enough for personal work. iTunes streaming worked without interrruption.
So the negatives with mobile internet are with software and not with the hardware. I did take the precaution of using a very long USB to mini-USB cable to keep the modem a couple of meters away. I don't think sitting within a few centimetres of an active modem all day long is an experiment I would like to perform on my own body.
What you (and I) really want is self-sufficient modem appliances which don't require the installation of complicated software or require multistep procedures to log in every time. Anything you add to your core OS just creates more potential problems.
And problems are time. And time is money. Alas, it appears the E960/E970 appliances can only be obtained from providers. Choose your provide accordingly.
OTHER RESOURCES
IT |
Tuesday, June 17th, 2008
I've come across what is purported to be the top Wordpress ping list over at SEOfeed.com. That list is supposedly regularly updated.
Here are the current contents:
http://blogsearch.google.com/ping/RPC2
http://api.feedster.com/ping
http://api.moreover.com/ping
http://api.moreover.com/RPC2
http://blog.goo.ne.jp/XMLRPC
http://blogdb.jp/xmlrpc/
http://coreblog.org/ping/
http://ping.blo.gs/
http://ping.bloggers.jp/rpc/
http://ping.cocolog-nifty.com/xmlrpc
http://ping.syndic8.com/xmlrpc.php
http://ping.weblogalot.com/rpc.php
http://pinger.blogflux.com/rpc
http://rpc.blogrolling.com/pinger/
http://rpc.icerocket.com:10080/
http://rpc.pingomatic.com/
http://rpc.technorati.com/rpc/ping
http://rpc.weblogs.com/RPC2
http://topicexchange.com/RPC2
http://www.blogdigger.com/RPC2
http://xping.pubsub.com/ping
http://api.my.yahoo.com/rss/ping
Brandon also recommends using the plugin Smart Update Pinger. Why? To make sure that when you edit a story and update it, that pings aren't sent out a second time. Some people (and it could include me) have had their Wordpress weblogs dropped from some ping services as spammers for repeated changes and new pings.
Strangely, Smart Update Pinger is no longer being supported - last version of Wordpress officially supported is 2.0. Daven in turn links to PingFix "for more functionality". Jan Piotrowski is quite explicit: "Pingfix is obsolete for Wordpress 2.1 and newer versions."
So why is Brandon recommending it?
Moreover there are two comments on his post which suggest that this ping list may be overkill. Taff suggests:
Just a thought Brandon but doesn't the use of Pingomatic mean that most of the sites are being pinged twice as nearly all are on the Pingomatic ping list? I thought I read somewher that IceRocket does much the same but can't find any info on their ping list. Just thinking about the possibility of spamming.
Joel asks:
do you think there is a chance that by pinging pingomatic and all the other services on the list that you may be sending pings twice to some services? one directly and the other by way of pingomatic.
Ken Xu mentions that four of the services just won't work for him.
For some reason Brandon Hall has no answers.
Two questions:
- Is this an accurate ping list?
- Do we need to tell Wordpress not to ping again on update?
I've like to have some answers.
WordPress |
Monday, June 2nd, 2008
I received yet another fake domain renewal notice, this time from Domain Submission Center. This one was from the city in which I went to university, Toronto.
What was new about this scam is that it used the search engine submission scam, within the domain renewal.
Go and learn more about the Domain Submission Center.
In line with this article, we've finally started our new tool, a directory of internet marketing scams to help people avoid getting ripped off.
Internet Marketing |
Sunday, June 1st, 2008
From the European Outsourcing Conference last week: it's difficult to do large scale outsourcing in Easter Europe:
The delegates realised that Eastern European nations do not have the scale of staff to resource huge outsourcing deals, with the majority (58 per cent) believing that different geographies are suited to different types of outsourcing projects. Only 17 per cent believed that Indian providers are leaps and bounds ahead of all other geographies.
Absolutely. There are just not the numbers here necessary to be able to do it efficiently.
What you can get are a few good programmers on a project.
Or a lot of reasonably multilingual (we're not talking about the Netherlands or Austria here) customer service bodies.
The very high euro - which seems to be carrying the local currencies with it - does make European outsourcing pretty much a local sport at this point. But there is no shortage of European companies and divisions looking to cut costs without transcontinental travel.
IT, Business |
Friday, May 30th, 2008
Have tech companies gone blue chip: no risk, little reward?
Over at purveyor of dubious business advice The Wall Street Journal, Mean Street says it is so:
The good news: Tech stocks are the blue chips of today’s economy. The companies are bigger and better run than ever before.
Still not convinced this sector has matured? Today, there are eight U.S. tech companies with market caps greater than $100 billion. Only three U.S. financial institutions are worth that much. Three. Last week, technology surpassed financials as the biggest component of the S&P 500.
The bad news: Tech stocks are the blue chips. Lower risk means lower reward. Are tech investors mentally prepared for the 10% equity return including a 2% dividend
Those are amazing numbers. Tech companies are bigger than banks. Curiously tech - and entertainment and weapon systems - seem to be the only products in which the US is a world leader these days.
Despite the huge market cap of the top tech companies, I think Evan Newmark is off base on the future of tech.
The Google IPO only took place in 2004. The tech market is as dynamic as ever. But there have been changes:
- The movement is away from hardware to software as a service.
- The movement is to profitable online models.
- Online advertising is about to really grow again.
- It's not that there aren't big rewards still out there in tech. It's just that it won't be in selling hardware.
The same thing happened when IBM released the PC, those cute old 286's. I owned one. Suddenly big iron and dumb terminals were of limited use. Companies like DEC hawking terminals and mainframes slowly just disappeared.
We are at the start of a new era of opportunity. The keywords here are:
- linux - open source
- flexibility
- interface
- availability
- low-cost
Anyone who gets those right - and we are multi year clients at least five of these companies (37signals Basecamp, Statcounter, Freshbooks, iContact, all highly recommended plus one legacy with poor service to remain nameless) with the average bill over $50/month - is in for ongoing paycheques. Web applications and commercial refinement for Linux may not be as exciting as multibillion dollar investments in new Silicon factories, Plasma screens or brand new OS's.
But the oceans are a lot deeper here. The opportunites are vast. The plays will be smaller - but clever VC's will have the opportunity to own half of a whole revenue stream which could be tens of millions of dollars per year.
For the moment, the VC's are making their money on acquisitions - and doing very well with the biggest sharks Yahoo, Microsoft and Google - circling in the water at all times. Some acquisitions are huge: YouTube at $1.6 billion (Google), del.icio.us at $7 million (Yahoo). But the current acquisition frenzy is just the top of the iceberg.
We are at the beginning of a paradigm change, of a real tech business renaissance, akin to the launch of MS-DOS in 1981 or the Apple Macintosh in 1984.
Note: This is not to say this tech-media renaissanceis without risk. Those who blow the pricing-access model can lose big in this new era, like the New York Times.
Business |
Tuesday, April 29th, 2008
Over the last month, the internet has been awash with stories of the fabled New York Times slow demise, as indicated by huge losses and impending layoffs. As the new Wall Street Journal hires, the New York Times fires.
The decline in New York Times revenue and readership surprised me somewhat, but I accepted the decline at face value. Given the New York Times atrocious editorial standards throughout the Bush regime, including aiming and abetting war crimes (Judith Miller), the loss in circulation seemed like just desserts.
But today, Comcast's list of the top 50 websites for March came across my desk. Based on unique visitors, guess who's at number 12 with 47 million unique visitors for the month? For refernce, that's just behind Wikipedia, Amazon and ahead of Facebook, CNET, Adobe, CBS and Craigslist: The New York Times Digital.
Comcast top 50 websites by unique visitors
Which set me to thinking what kind of second rate media (advertising sales) strategy would it take to lose money with the number 12 website.
I don't know how reliable the Comcast numbers are so I went and ran some quick checks on Alexa.com.
While the numbers are not as good over at Alexa (the New York Times is number 25 in the United States), the reach is still fabulous. Where the New York Times falls down is in page views. They average only 3 page views per unique visitor.
Contrast those page views with Wikipedia at 5 page views per unique visitor or Google at 7 page views per unique visitor or Facebook with 25 page views per visitor.
nytimes vs wikipedia vs amazon vs digg vs facebook
Why are the page views so low for what should be a multiple page view experience (who opens a newspaper to read one article)?
The silly registration hoops. You can't read NYTimes.com without jumping through hoops. Either you register with them (and they kill off the registrations occasionally, requiring reregistration as I did register once but my login is dead) - or you can't view the site. They've taken to effectively blocking BugMeNot.com which used to be the best way to log in to NYtimes.com in a hurry.
NYtimes registration firewall lockout
It is possible to bypass the login with a very clever automated NYTimes registration bot which generates a totally anonymous profile in a nanosecond. You can link to stories with a special URL which can be generated with a javascript bookmark.
All of these shenanigans are time consuming and annoying. And guess the results - you arrive at NYtimes.com from a weblog on a permalink. You seem something else interesting and want to go and have a look. Paff - you're locked out. Register or forget it (correction: apparently you do get four or five peeks before you hit the registration firewall).
Once you do register, a good number of the stories still won't be accessible without paying a ludicrous fee ($5 for a newspaper article?).
So now the business model is to start letting journalist go. So less high quality (I think they got rid of Judith Miller, no? and lately they've been letting a lot of Republican cats out of a deep dark bag - so let's be optimistic) original content. Less information for the search engines, less possibility of generating original content.
I don't understand these guys. Their assets are:
- brand
- original content
What they think are is their third asset - traffic/eyeballs - is only built on the first two. So when they dump a certain amount of the original content, down goes the traffic. The only thing differentiating the NYtimes from all kinds of news and aggregators is their pressroom.
The first attack came on the quality - the brand:
Well, they didn't let Judith Miller go. And her fact-impaired cheerleading for the Iraq war has helped land us in a mess that's going to last at least another 10 years, in my opinion. And kill several thousand more US soldiers, several *hundred thousand* more Iraqi civilians, cost us trillions more dollars, and worse...
The problem the Times has isn't the quantity of reporters, it's the quality. People know they can't rely on truth from the times on important issues, because they *have not* been able to rely on,
Wave two is coming on quantity (number of reporters).
With the traffic, search engine rankings and original content they have, someone in the NYtimes media department is doing a seriously bad job of promoting and monetizing the site.
Years ago Salon.com came up with the model of subscriptions or day passes which require viewing advertising. It's still working for them (and they are looking for New York advertising reps). In the meantime, couldn't somebody clever implement this or come up with something new?
If the New York Times media department doesn't hurry up and figure out how to extract revenue from their brand and original content, the next loss will be traffic. From the loss of traffic, there will be no return.
This is not a critique of their tech department who seem to know what they are doing:
- high search engine rankings
- good internal search
- attractive layout
- open source contributions
But rather a critique of the business side of operations.
WordPress |
Monday, April 28th, 2008
For years, I've been on the Site Build It list. SBI is the creation of the rather annoyingly gushy Ken Evoy who never stops his carnival barker cries about his one-stop-site-creation tool.
Ken Evoy Pumping Site Sell
Evoy's been at it since the bad old days when the internet was a mess and Site Built It! did have the advantage of actually getting a website up in some form - easier than coding html from scratch for the neophyte.
Throughout SBI's history, Evoy has shrieked about his process and his proprietary tools. On the surface, a clear process and proprietary tools are a good idea. Probably worth the price of admission (or so I thought at the time). The issue with the proprietary tools (which otherwise might be a good deal) is that you can only use them a little bit. Come and play for one hour per week, see you next week. Not exactly inviting brainstorming or creativity.
In contrast, the indepdendent expensive (many of which are free) tools Evoy condemns let you use them as much as you like once you find them.
Over the years, I've learned not to expect much from Evoy's newsletters (sometimes for six months at a time, they get relegated to the read later bin). Still it's worth sometimes checking in on somebody who's multiyear obsession has been selling ecommerce sites. Another perspective.
In the last couple of years the internet has changed and it's actually quite easy to put a website up. Just buy a hosting account (a single domain account is $3 to $7 Ken, not the $10 to $15 you cite), click the one step install button and you have vanilla Wordpress (or Mambo or Joomla or whatever else catches your fancy). Or pay nothing and sign up at Wordpress.com and have a better than vanilla Wordpress install with lots of attractive themes ready and waiting for you and an active forum.
The ease of putting up a high quality website - almost all of which look better than Site Build It websites and are easier to post to - is naturally a huge threat to the SBI business. Why pay Ken Evoy $300 per year per website for hosting which should cost $50?
Evoy's latest missive starts yet another hysterical title "Why blogging is a massive mistake!" Exclamation mark is his.
Writing a weblog is not a massive mistake. Handled properly, a weblog does wonders for your website traffic and search engine standing. But taking away the hype, this time Evoy does have a worthwhile point about weblog type sites (Wordpress in particular) - i.e. they date like stale newspaper. I can confirm the tendency from my own sites.
By publishing a weblog, you are effectively creating just a daily news source.
What happens if you publish a very good article which has value as a permanent reference? It stands alone in your weblog. People come, read the single article and leave. There may be other interesting content on your weblog for them to read but the visitor can't be bothered to ferrret it out. If your writing or content is extremely compelling, perhaps some visitors will read a certain amount of your content. But then they will leave. Which quite frankly for an online journal is fine. You're not selling anything.
But for a business, this isn't so good. What you want is to create an information resource for people in your business, which will bring them back again and again. An information structure which invites them to find immediately the other relevant areas of interest.
And Evoy quite correctly points out that this is the built-in model for Site Build It:
Blog posts are created and stored in chronological order. A good blogger will produce a post that is useful today, but who will read it in three months? Even when bloggers go to the extra effort of archiving their posts by "keyword categories," the articles are dated and not rewritten into coherent definitive articles. Usefulness plummets with time.
How does a Theme-Based Content Site differ? Instead of a stack of old newspapers, each resembles a good resource book about its theme, composed of useful, original articles ("Web pages") that cover related topics in some depth. Written in each small-business owners's unique voice, and based upon that person's experience in the field, they are useful resources that visitors return to over and over.
Evoy correctly points out that a photography weblog would just be one in a million, posting the nattering about the latest cameras and software:
How would a blog be presented? A stream of disjointed photography tips would be organized by "date of post." And posts on any given topic (ex., "portrait lighting") would be separated by time (weeks or months apart), each covering only a certain aspect of the topic. On the other hand...
Definitely not the right one to pull someone into your website. Evoy contrasts the above weblog site with this siloed sitemap for a static site:
site build it silo site
This time Evoy's absolutely right. Someone looking for information on photography lighting would gradually be led through the whole of your website, would bookmark it and come back as a reference. All of this assumes of course that your content is top-notch (and Ken, let's be frank, there's not too many people capable of creating top-notch content, on or off the SBI rolls). But with a static site structure at least you stand a fighting chance of retaining your visitor and becoming a reference.
In any case this is a huge insight. Pages instead of posts something I've been playing around with in the static pages section in Foliovision. Our client sites are also largely hierarchical with the weblog performing weblog functions (added value).
What I've been doing is making a static page instead of a post and then publishing a small announcement on the weblog section.
Unfortunately some of the news outlets which republish my content will not link to static pages or to articles which are more than 24 hours old (a pain in the neck, as after publishing a major article I like to come back to it 12 hours later to proof it and add or correct illustrations).
Going forward, I am going to build up the static pages sections very actively. When I first publish a post, it will go into the weblog, but within a few days. There is one small issue which is comments. We enable comments on pages so visitors will still be able to comment on the static page. But often some of the comments come in right away (on the weblog version).
- Do I leave the comments on the weblog post or move them to the static page?
- If I choose to move the comments to the static page, there is no mechanism to do so inside Wordpress. We'd have to build a plugin.
BTW, this sort of question is what you are paying Evoy to solve for you with either no solution (in this case) or his solution. For an inside the box thinker (or someone with very little design sensibility and/or minimal interest in technology), SBI solves a lot of problems. For an existing six-figure business, there are better ways to bring your business online than SBI DIYism. I do agree with Ken that business owners should have better things to do with their time than spend it troubleshooting websites or optimising their sites for Google.
If you're interested in having a closer look at the Site Build It system and way of thinking, Ken Evoy offers a number of free ebooks on writing for the web, selling services and montization. SBI's claim ithat the free ebooks are better than a lot of the pay ebooks out on internet marketing is more or less true. Given the rubbish sold as ebooks that's not necessarily saying a whole lot. The link above bundles several of them into a single zip file for your convenience.
Personally, Ken's writing style drives me up the wall (he's been described as rah-rah), but the bulk of the information is good. I just can't read past his marketing speech. The formatting is bizarre as well. I wish the guy would hire a graphic designer at some point. Why does he write Sidebar and then not make the sidebar a sidebar but whack it right into the middle of the text?
Some other references
Internet Marketing, WordPress |
Saturday, April 26th, 2008
Just ran into a time consuming hiccup trying to work with a Clone CD Image. I hope the rather detailed explanation below will help someone else deal with a Clone CD Image faster on his Mac.
I'd downloaded a 600 MB disk image to use with Parallels. The disk image came wrapped in a .rar format.
Unpacking the .rar file was very difficult.
- StuffitExpander crashed
- Forklift couldn't handle it
- Archive Utility just wanted to compress it further.
Finally I was able to unpack it with IAarchiver, slowly but certainly.
After unpacking, I was offered a directory with three files in it:
- image.img
- image.ccd
- image.sub
Clone CD img ccd sub files
I had no idea Windows used the archaic Mac .img format. It turns out that Windows doesn't. What I'd inadvertently downloaded is what is known as a Clone CD image. I was unable to mount it on my Macbook. No known solution or software for Mac OS X can handle Clone CD images. BIN and CUE files are no problem, but CCD files - there's nothing out there.
- DiskUtility wouldn't do it (error mounting the disk)
- Toast 6 wouldn't do it (although Toast did make me a coaster with the file when I changed the .img extension to .bin and used Copy Disk Image)
- Burn also made me a coaster, even after I followed elaborate instructions on renaming files and creating a fake .cue file
- Firestarter was a non-starter. It didn't want to use my DVD burner at all. In any case, Firestarter seemed confused about what to do with the files.
If one had a Windows computer with a CD burner nearby, the logical step would be to burn the .rar file to CD, move it to the Windows computer, mount it with the freeware Virtual Clone Drive (download) and burn it as a disk and bring it back to the Mac as a ready made CD. I didn't have a Windows computer handy - and it seemed pitiful that my Macbook with 10.5.2 couldn't handle a foreign disk image format.
I managed to convert the Clone CD .img / .ccd files to an .ISO file in the end though.
How?
Command line. Those command line addicts over on the Linux/BSD side of the fence have written a lovely utility called ccd2iso. It's a single command.
Unfortunately, ccd2iso is not included out of the box in Mac OS X.
You have to install MacPorts (ccd2iso might also be in Fink). If you haven't got MacPorts installed it's a bit of a drama.
First, you need Xcode. Xcode is the developer toolset.
If you install MacPorts without Xcode, you will be able to install it, but when you try to update you will get nothing. With DarwinPorts you will get this error:
sudo port -d selfupdate
DEBUG: Error installing new MacPorts base:
shell command "cd /opt/local/var/macports/sources/
rsync.macports.org/release/base &&
./configure --prefix=/opt/local --with-install-user=root
--with-install-group=admin
--with-tclpackage=/Library/Tcl && make && make install"
returned error 1
Command output: checking build system type...
i686-apple-darwin9.2.2
checking host system type... i686-apple-darwin9.2.2
checking target system type... i686-apple-darwin9.2.2
checking for sw_vers... sw_vers
checking Mac OS X version... 10.5.2
checking MacPorts version... 1.6.0
checking for gcc... no
checking for cc... no
checking for cl.exe... no
configure: error: no acceptable C compiler found in $PATH
See `config.log' for more details.
So install Xcode as you were supposed to (I have Xcode installed on another Mac and thought I had it on the Macbook as well).
Once you have Xcode installed, first you need to bring MacPorts | DarwinPorts up to date:
sudo port -d selfupdate
This takes a little while and generates a lot of messages: you are installing about a one thousand open source components, including Gnome, KDE and X11. I ran the updater a second time just to make sure that everything was caught the first time.
You know you're okay when you get this message:
The MacPorts installation is not outdated and so was not updated
DEBUG: Setting ownership to root
selfupdate done!
The next step is installing ccd2iso.
sudo port install ccd2iso
And to convert the rogue .img to .iso:
ccd2iso image.img mycd.iso
The conversion takes about ten minutes with a 600 MB file, but I can confirm that my disk image works as a perfect self-contained ISO. The .iso can be mounted in the Finder with DiskImageMounter and can be used directly in Parallels.
ccd2iso finished ISO
If you've already got Xcode and MacPorts installed, you're golden. You are two minutes away from converting your Clone CD image to ISO. If not, you have a 1 GB download ahead of you (Xcode) and some tinkering.
I wanted to have access to MacPorts on this computer in any case. An open source drag and drop converter for the Clone CD format for Mac OS X might be nice though. It would have been faster!
More discussion
emuforums: pscx
IT |
Wednesday, March 26th, 2008
We've been properly labelling and tagging our images for years. Some of our websites get most of their visitors from Google Images.
Google Images is the greatest SEO reserve left in the world. Chris Silver Smith of Netconcepts let the cat out of the bag in 2006 and told the whole world about optimising for Google images. But it's hard work optimising images for Google Images and most webmasters still can't be bothered. There's still gold - or at least visitors - in those hills.
As Chris didn't cover the technical details in-depth, here's a step by step guide for optimising your images for Google images.
Most websites publish their images like this:
<img src="/images/192a/986943.jpg" alt="image">
Where's the problem? Missing height and width, meaningless directory name, meaningless file name, generic alt tag.
Here's what a properly formatted image should look like:
<img src="http://foliovision.com/images/2007/08/zen-fanless-power-supply-400.jpg" alt="Zen Fanless Power Supply" width="400" height="340" />
For bonus points link that image to a larger version of the same properly labelled image:
<a href="http://foliovision.com/images/2007/08/zen-fanless-power-supply-big.jpg" rel="lightbox"><img src="http://foliovision.com/images/2007/08/zen-fanless-power-supply-400.jpg" alt="Zen Fanless Power Supply" /></a>
For extra bonus points put that image in a h5 tag with a proper caption, close to if not identical to the alt tag:
<h5><a href="http://foliovision.com/images/2007/08/zen-fanless-power-supply-big.jpg" rel="lightbox"><img src="http://foliovision.com/images/2007/08/zen-fanless-power-supply-400.jpg" alt="Zen Fanless Power Supply" /></a><br />Fortron PFC ZEN fanless power supply</h5>
If all that sounds like a huge hassle - when you do it for every image - you are absolutely right. It is a huge hassle to optimize for Google images by hand.
Which is why we built the SEO Images (part of Foliopress WYSIWYG) plugin.
With SEO Images, all of the above is happens automatically.
You only need to give the image the correct name (words separated by hyphens) and upload to the correct directory.
Automatically all the rest is added to your image:
- alt tag
- thumbnail (whatever size you prefer)
- link to larger version image
- caption
- width and height
- lightbox
If you want a lot of visitors from Google Images, you only need to use SEO Images for a few months and you will have the rankings and the visitors to go with them.
Here are the Google Images result for our example from above, the Zen Power Supply. Of 107,000 images, spots one and two are from Foliovision.com. The large and the small version of that image.
SEO Images Google Images results
Why a few months? Historically indexing in Google Images is much slower than for the rest of Google.
Chris and Stephen, in the future, please keep our secrets to yourselves!
Also check out Problogger Formatting images for SEO.
SEO |
Thursday, March 13th, 2008
Over at LifeHacker a fascinating discussion of monitor size and productivity. It caught my eye as I've recently moved from a 30" monitor to a 24" and 12" setup. Strangely I find I'm more productive on the second setup.
In any case, people have all kinds of strange setups including one guy with six 15" LCD's all on a special mount. I think he's onto something. As I said, I've gone from a single 30" to a 24" (1920 x 1200) plus a 12" (1280 x 960). At home I now have 20" (1680 x 1050) plus the same 12" (Macbook).
I stopped running the 30" as my Macbook can't do Dual-Link DVI.
I thought my productivity would go down. No way. Substantially up. Managing the windows and flipping between applications was a hassle on the 30".
With a dual monitor setup, all the distractions on small monitor. Work on big monitor.
That said, I much prefer the 24" as a main monitor. I am less productive by an order of magnitude except when web browsing and writing on the 20". It's just not big enough to handle two full size documents (without having 8 or 9 pt antialiased type to squint at). 24" is the sweet spot.
simple dual monitor setup 24" and 12" (macbook plus 1920 x 1200)
Finally, if you can avoid TN screens on your main screen. Sometimes you want to stand up and look at your work. Sometimes you want to lean back and look at your work. You can't do it. The colours go all wonky. Things get dark. The monitor distracts.
Caveats: the 30" monitor does have two useful functions still. It is superb for sorting and editing high res digital photos (3000 pixels and up). It is also superb for running FCP. As I don't do any video editing anymore only the first still applies to me. When I bought the 30" I was still planning to edit video. Running a company leaves no time and takes no prisoners though.
IT |
Sunday, February 3rd, 2008
I've been wondering about Wordpress plugin overhead for some time. How does one keep track of how much processor time and overhead any given plugin requires?
We run fairly streamlined Wordpress installs at Foliovision with about 30 active plugins per site. A lot of them are one-trick ponies developed in-house so we know the code isn't creating a huge load.
But anyone who has been working on Macintosh computers from the old days (System 7, 8 & 9) knows very well that every extensions (and some people were running 50 of them) slows down your compuer and increases the chances of a system conflict. There were whole expensive utilities devoted to keeping extensions and control panels under control. Any one else remember long hours spent with Conflict Catcher?
conflict catcher screenshot
Here's what WPdesigner.com has to say on his own plugin issues:
With the WP Download Monitor plugin, the front page of my blog had to operate with 136 queries on every page load. After uninstalling that plugin, the front page needed only 10 queries to work. 136 versus 10 and all I have to do is give up tracking the downloads, hmmm.. oh what, oh what should I do? I deactivated / uninstalled it, of course.
So some plugins are clearly worse than others. I hope I didn't publish that post recommending Lucia's LinkyLove for DoFollow. In the comments to the post above David Airey notes.
I found the real culprit - the LinkLove plugin for DoFollow comments.
By de-activating it I’ve dropped the queries from 117 to 24 on my blog post with 46 comments.
Lucia's LinkyLove plays complicated games rewarding commenters for more comments with better links. A pox on this sort of brownie point system. I will get DoFollow turned on and will rely on Akismet and common sense to keep the spammers off. Everybody else is welcome to followed links.
Ironically enough David's post recommending Lucia's LinkyLove still stands and is ranked high - that is not taking very good care of one's visitors. Perhaps he just forgot. Even more ironically, Lucia publishes another plugin to track CPU greedy plugins. I don't recommend this plugin, as it is ugly and is less useful than adding a simple PHP tag to your footer to track both database queries and CPU time.
<!--<?php echo $wpdb->num_queries; ?> <?php _e('queries'); ?>. <?php timer_stop(1); ?> <?php _e('seconds'); ?>.-->
At this point you can load any page on your weblog and check the number of database queries and the CPU time just by viewing source. If you are really on a roll, you can remove the html comment markup from around the PHP tag and you will be able to see the load times and database queries without checking source. So will your visitors, but for a half an hour of quick testing it really doesn't matter.
Here's the PHP code alone:
<?php echo $wpdb->num_queries; ?> <?php _e('queries'); ?>. <?php timer_stop(1); ?> <?php _e('seconds'); ?>.
So with 30 Wordpress plugins running, what's the damage on Foliovision.com and client sites?
wordpress database queries cpu time
On Foliovision.com,
How important is it to avoid greedy Wordpress plugins? If you get little or no traffic or you are on a dedicated server, it's probably not all that important. But if you are on shared hosting and you ever get Dugg, Slashdotted or listed at Yahoo, watch out. Your hosting may be disabled for a week due to the one or two errant plugins.
Here's a small case study of a plugin gone bad: Comment relish issues #1, #2 & #3. This one has a happy ending as the plugin was eventually repaired and made useable.
WordPress |
Thursday, January 24th, 2008
I've just been debugging display issues in the CSS in Knowlege Constructs FAQ-Tastic tonight. Firefox and Safari on Mac were a breeze to get right: just pull all the margins and padding off of ol.faq with a .nonumbers ol class that I'd already been using. It was especially easy to figure it out with the Web Developer's Toolbar on Firefox.
Unfortunately a quick excursion over to the Darkside and Internet Explorer (the blinkers through which 92% of the visitors to our clients still see the web - among Folivision vistors Internet Explorer users are a minority), showed that the CSS code just wasn't working. Indentation had gone totally astray.
In the absence of Web Developer's Toolbar for Internet Explorer, there is no way to get instant Internet Explorer preview. The closest thing is to open up the file directly from the server and save it back to the server. Usually, I am set up with two monitors on my desk, a 20" Samsung 205B for the Windows box and an HP LP3065 for the Mac work station. It's just a matter of editing in CSSEdit or BBedit on the Mac, saving onto the server and pressing F5 on the PC keyboard.
We've installed a Linux machine now - the first of many - and I had to give up my 20" Samsung 205B and plug the Windows box back into the HP LP3065. (Both monitors are highly recommended, btw.)
Pushing input and switching keyboards was not efficient (3 movements instead of one, along with a screenflash each time).
So I decided to take the plunge and go looking for a Windows XHTML/CSS editor which would allow me to open up files from the server. It was either that or move a monitor.
html kit screenshot
I'd had a quick run-in with HTML-Kit a couple of nights ago which I found via somed SEO research I was doing (htmlkit are doing some serious link selling) but had not been happy at all with the tool. It was ugly and clumsy. Nothing like being at home on BBedit (which while arguably drab, is not clumsy). The website was particularly stressful with it's ugly and unreadable four column layout. Would you want to trust your html and CSS editor to people who can't build a readable web page? Me neither. While version 292 is free, all future versions and advanced functionality are relatively expensive, with just part of the pro package costing $65. I don't know if the guys at htmlkit have a drug habit they are supporting with their newfound commercial activities and advertising but something is seriously amiss.
So I decided to look more seriously this time. After reading quite a few reviews, I managed to find what is the finest development environment on Windows fairly easily - Aptana Studio. Our programmers here at Foliovision work in a cousin of Aptana called Eclipse and they love it. The website itself is gorgeous. I was sure I'd found my own Windows web development tool.
But when I downloaded Aptana Studio I was shocked to see 90MB come down the pipes. There is no way that I want a 90MB text editor. Not for my HTML and CSS editing. I still believe in lightweight software (speaking of which Open Office, it's time to go on a diet - you're looking too much like Microsoft's own products).
UltraEdit had come recommended by a friend who is really into ugly but efficient Windows apps. UltraEdit may be efficient but I won't be finding out as the program is so hideous that there was no way I wanted to even trial it. Weighing in at 10MB, UltraEdit did pass the efficiency scale. What particularly turned me off was the aggressive marketing selling tools which should be included in the editor itself (UltraCompare for file comparison) or of no particular use to someone looking for a text editor (UltraSentry, a Windows web history eraser). Why do I want to spend $139 to get a text editor?
ultraedit screenshot
EditPlus is even uglier than UltraEdit and nearly as expensive. In 2008, it should be a crime to write applications so hideous and charge money for them.
On the same website, where I had found my UltraEdit review, I found another freeware text editor reviewed: HateML. Apart from the silly name (comes from German pronunciation of HTML - ha -T-M-L), the feature list looked good.
- syntax highlighting for xhtml, css, php and javascript as well as some other languages I don't care about: it's much quicker to read coloured code
- automatic syntax checker: nothing like the coloured code for showing you typos before you even upload the file to the server - preemptive debugging
- code hinting, just like with CSSEdit. I actually don't like code hinting much as I know the tags for the languages I use and find that it gets in the way most of the time
- built-in FTP client with server side editing (not upload and download)
There are some other nice features which mean nothing to me but would be very useful in the right hands, like PHP debugging, site management, website templates, MySQL managers. You won't outgrow Hateml too quickly.
hateml screenshot - nice and simple
A quick visit to the website showed that Hateml's programmer Michal Gajek cares about aesthetics and cares about code which validates. A good start.
michal gasek web standards
The application itself is a delight. I was able to use the big monitor to good advantage editing the CSS in one colourful window and with a quick click and press of F5 see a preview directly in Internet Explorer. The ftp upload is even faster than on my Mac. Strangely enough Hateml is written in Delphi. This is the first usable Delphi program that I remember using. I much prefer the Delphi look and feel at this point to a Java UI.
It might be that if I spent a lot more time on the Windows side, I'd want a full fledged environment like Aptana but for now, I'm delighted with Hateml and thoroughly recommend it as one of the unsung heroes of the web coding world.
It really appears that the main thing many commercial applications - UltraEdit I'm looking your way - have going for them is hype. If I'm Hateml is still doing service for me in three months, I will donate. I recommend to everyone to donate to their favorite freeware authors, after a three to six month trial. There are a few freeware programs which I have donated to three times over the course of years. Those donations leave us choices.
In case you still remember about the FAQ styling, here are your Toronto home buyer's questions.
IT |
Wednesday, January 16th, 2008
I've just deployed a FAQ page onto one of the client websites using the Wordpress FAQ-Tastic plugin onto one of our client. I took some notes on the deployment and have written up a long FAQ-Tastic review. FAQ-Tastic was written by John Godley for Knowledge Constructs.
FAQ Tastic's default layout
Executive Summary
While not without issues, if you are thinking of using your FAQ page as a live resource you should definitely consider FAQ-Tastic. There's a lot of power under the hood. Don't miss the full FAQ-Tastic review with deployment notes.
WordPress |
Thursday, January 10th, 2008
Photoshop CS3 Save to GIF Droplet
(Mac OS X version)
If you use Photoshop CS3 and post screenshots to the web, this little droplet will save you a lot of trouble. For some reason it is impossible to convert ImageReady or previous Photoshop versions Droplets to Photoshop CS3.
Installation and Usage Instructions:
- download the zip
- decompress
- move to the folder of your choice (I have a special folder for Photoshop and Image Ready droplets)
- title your images for upload (spaces are okay - PS3 will convert them to hyphens)
- drop your images on the droplet
- your web ready GIF's will appear in your desktop folder
For equally unknown reasons is also extremely difficult to create a droplet which will actually open your image and resave it as you would like right in the folder where it lies.
Even my version here will save the GIF file to your desktop, rather than the folder where the original lies (my preference). Desktop isn't bad, as you can then upload the image and archive the extra desktop files every couple of days in a date named folder in a desktop archive folder.
Current advice on the Photoshop forum is to use the Image Processor or to move to Fireworks (a program I and many other Photoshop users have never opened up, let alone want to leave running constantly in the background). Image Processor is a nice piece of kit, but it's suited to bigger jobs, whole folders. What I need is something to convert my screenshots from .png to .gif or .jpeg for posting to the web. Nothing against PNG - it's a great format, but my web server will be very full very fast with the 500 KB files it generates. Moreover, much as I like images, there's no reason for anyone to have to wait that long to download screenshots.
So I need a droplet just to take the PNG and save it as GIF. This is that droplet.
BTW, you should never convert your screenshots to jpeg, except a very high quality compression algorithm. The only reason to prefer jpeg to GIF for screenshots - which are usually mainly text - is if they include photos.
References:
WordPress |
Tuesday, October 7th, 2008
Uncategorized |
