The Best Long Introduction to a Company and a Product

On Monday Asana had an open house publicly revealed the state of the project management web service they are developing. Justin Rosenstein did an amazing job presenting the vision of Asana and doing a demo of the Asana product.

I don’t think I’ve seen a better long introduction to a company and a product. The talk is about 45 minutes, and Justin maintains his mellow. He is articulate, cheerful, and on point throughout.

Emerging Web Fonts

“It turns out that Mobile Safari doesn’t support as many embeddable font formats as the desktop version, so Google sends an SVG font version to iPhones, iPads, and iPod touches, or anything pretending to be them. And it looks like Google’s SVG fonts contain only ASCII characters, while the other formats have full character sets.”
Derek Miller, “A weird Safari-Google bug”, June 21, 2010

I haven’t encountered anything like this, but I found it to be an interesting interplay of emerging technologies.

It reminds me of the headaches of different database encoding and WordPress.

Ubuntu Linux Still Searching Google

Ubuntu has flim-flam-flapped back to Google for search.

Obviously, I think this is a good move. To recap switching to Yahoo for search would have alienated users because it’s a worse search engine, but more important would have overwritten people’s existing experience on upgrade.

In an email to the ubuntu-devel mailing list titled “Follow up to Default Search Provider Changes for 10.04Rick Spencer writes:

Each release we determine the best default web browser and the best default search engine for Ubuntu. When choosing the best default search provider, we consider factors such as user experience, user preferences, and costs and benefits for Ubuntu and the browsers and other projects that make up Ubuntu. Up until Ubuntu 9.10 these defaults have always been Firefox and Google. Earlier in the 10.04 cycle I announced that we would be changing the default search provider to Yahoo!, and we implemented that change for several milestones.

However, for the final release, we will use Google as the default provider. I have asked the Ubuntu Desktop team to change the default back to Google as soon as reasonably possible, but certainly by final freeze on April 15th.

It was not our intention to “flap” between providers, but the underlying circumstances can change unpredictably. In this case, choosing Google will be familiar to everybody upgrading from 9.10 to 10.04 and the change will only be visible to those who have been part of the development cycle for 10.04.

All thanks the Ubuntu Masters!

One Experience

After reading Mark Pilgrim’s dynamite “One” (spoiler: there is only one of you), I came across another of his excllent articles from a month ago, “Ubuntu and Yahoo“.

Don’t Downgrade the Experience

He quotes Rick Spencer, Canonical:

No, this will effect [sic] upgrades if the computer is currently set to Google. This is not because of anything special for this particular change. This is because Ubuntu always changes to new defaults for users who are on old defaults.

Once I’ve installed software, it’s no longer “defaults”, it’s my experience. Any change to the experience better be to improve the experience, otherwise you jeopardize alienating me and your other customers.

It Can’t be About Money

The search engine in Firefox is Google because it’s the best (of the infants).

In my co-worker Noël Jackon’s post today “Art First

I respect Jay-Z for his music, but love the man for his words.

via YouTube - “NY-Z” – An ABSOLUT Collaboration with Jay-Z.

At the Jay-Z saysat the 11:30 mark “it can’t be about money. There has to be something in there that is true to both sides. … when the align … At the end of the day no one loses when it’s like that.”

Stay true to yourself, and true to your customers.

I’m Staying Hopeful

Ubuntu has hugely help make Linux and open source flavors many more people can enjoy.

I’m staying hopeful that Ubuntu will stay focus on their customers’ experience!

In the meantime, I’m also staying pragmatic and swallowing bitter proprietary medicine everyday.

Having Fun

My co-worker Alex “Viper” today shared this parody of Jay-Z “Empire State of Mind” titled “Galactic Empire State of Mind” by College Humor:

Customers’ care about one thing, a good experience.

April 8th Update: Yahoo will not be the default search engine in Ubuntu 10.04 Lucid Lynx. Related article “Ubuntu Linux Still Searching Google

Built It or They Won’t Come

‘… Most of the successful versions of HTML have been “retro-specs,” catching up to the world while simultaneously trying to nudge it in the right direction. …

That’s not to say that all shipping code wins; after all, … Code is necessary but not sufficient for success. …

The ones that win are the ones that ship.’

Mark Pilgrim, “Why do we have an IMG element?“, Nov 2nd, 2009

The article includes fascinating excerpts from some of the leaders in this almost-17-year-old work in progress, HTML.

SEO Experts and the Complex Web

This could go on for days:

SEO Guy: I do x, y, and z, how is that bad?
Derek Powazek: X, y, and z isn’t SEO, it’s making good websites.

SEO Guys, please change your titles to Good Website Makers. I don’t care why, just do it. Derek Powazek insists.

Posted by Jason Kirk on 12 October 2009 @ 3pm

SEO Blocks

SEO Blocks by Vishraval. Wikipedia Hosted. PD Licensed.


Is the very witty comment by Jason Kirk on Derek Powazek’s “Spammers, Evildoers, and Opportunists” published October 12th, 2009.

Derek’s intense, absolute position and lack of a shared language with SEO Experts leads to little opportunity for real discussion between the parties, but at the heart of Derek’s article is a sentiment shared by many web developers: we are extremely frustration that there continues to be a market for search engine manipulation.

This reminds me of Cory Doctorow’s “All Complex Ecosystems Have Parasites“, published 16 March 2005. It makes me think that I’ll happily pay the price of having to put up with SEO Experts, if the alternative is a less flexible and diverse Web.

Export, the Second Feature

Quartz by Mikael Hvidtfeldt Christensen. Flickr Hosted

Quartz by Mikael Hvidtfeldt Christensen. Flickr Hosted

I used to joke that the second feature to write is export. I don’t joke about it any more. Export is the 2nd feature you should implement for your software or web service.

There is nothing that says you care about your customers like making it easy for them to get their content out. Bonus points if you choose an export format that is already popular and well documented.

If you really love your customers, the exported data will be richer than the raw material they originally entered.

That, of course, makes import the 3rd feature to write. Don’t support importing from competing applications until your product is ready, because migrating from another product is already a scary enough situation without finding yourself using a buggy, incomplete product.

5% of Nothing

More than 5% of Nothing

"Sync Alert" Adding 1 Contact with iPhone OS 3.0 with Mac iTunes 8.2.1 (6)

Makes me laugh and cry a little.

This alert is likely meant to warn that a whole lot of data is being added, modified or deleted as part of a sync. It’s an “oh crap, likely either you are doing some wrong or the software is”.

I’m intrigued that UI Expert Aza Raskin (Humanized & Mozilla Lab) finds this alert the “The first good use of a warning I’ve seen!”

I’m not confident that it is generally helpful. Reading online, it does seem that Mac Sync has been quite buggy historically, so this would likely have been very helpful, but does make me nervous that this alert is a bandaid instead of the needed medicine. I would be interested to find out the use cases, and the scenarios where this has been needed.

Synchronization of data between two (or more) sources is a really hard problem. Well the hard problem is mostly related to conflicts when something is changed in two or more places. Daniel Jalkut once wrote, ‘Every developer faces the decision: “Do I want to be known as the jerk who won’t implement sync, or the jerk who can’t.”‘

I’m not actually syncing, because I don’t use Notes on the Mac (does it exist?). It’s really just doing a backup.

In this case, that is the “Sync Alert” of syncing an iPhone running iPhone OS 3.0 with my Mac over ethernet to iTunes 8.2.1 (6). I’m syncing 1 note.

Assuming that this behavior is generally useful, that I’m receiving a warning when 1 note is being added brings up the most obvious issue. There should be a minimum threshold before this exception behavior is triggered. It shouldn’t be 5% of nothing.

When you’re just starting to sync with your Mac, this could be a fairly high frequency alert. It depends on how quickly you add items; how quickly you get to more than 20 items in a category. You’ll also see this alert again when you start using a new feature (new type of item). This leads to the 2nd issue, the alert and warning language — I’m not seeing any yellow, but I’m sure feeling it. Because it is potentially high frequency, it should be presented and worded as a friendly confirmation.

The 3rd issue isn’t obvious from this screen shot. Another clue that it should be a confirmation is that sync does not continue until you have cleared this alert. This is actually problematic, because the alert is non-modal, meaning you can hide it or bury it under other windows. The worst part is iTunes with it’s animated progress bar makes it look like the sync is still progressing. This alert should at least stay on top of iTunes.

The 4th is I don’t think add, modify, and delete are equal. Adding an item is an easier event to undo then a modify or delete. I can just delete it. If something is modified or deleted, it may be hard to recover the lost information. The different events should have different weights. Adding an item should only trigger an alert if a lot are added.

How many is a lot? That brings me to the 5th and final issue that readily comes to mind. It likely shouldn’t be a percentage at all. How long would it take before 5% is a really big number? Probably not long at all. If I have 1000 business contacts, a sync would have to delete 50 of them for me to be notified. Again assuming this alert offers some protection, by using a percentage, even as low as 5% percentage you are penalizing your most passionate customers. Possibly, it could be a percentage that also factors in things like amount of time since last sync or what operations resulted in the changes, but that would likely get complex quick and I suspect the developers would more quickly find the sweet spot by using a constant number (possibly variable on a curve. I hesitate to recommend user defined variable).

Later in that discussion by Aza linked above, he does add “It could be laid out better, but I like the idea of “uhhhhh, that’s dangerous”. Of course, undo is always better :) ”. I don’t think undo would be the silver bullet. I definitely think there is value in confirming changing a large amount of data particularly when the change is destructive, and the need to undo might be overlooked. I just don’t think Apple has polished this implementation.

I feel a bit like I’m playing dirty writing this article now when the next version of Mac OS X, Snow Leopard is only a couple of months away. Still, I was hoping this would be resolved with the new iTunes for iPhone OS 3.0. I’m still hopeful that Snow Leopard with it’s attention to polish might surprise me here.

Gmail’s Opportunity to Help Protect Against Tagged.com Mistake, Spam, and Phishing

Now for the part of the Tagged.com story, I really wanted to tell. As I mentioned in “Tagged.com Spam? Phishing? Nice Guys? My Personal Story” I try to look at situations and problems from different angles.

There is a clear opportunity for online email providers and social networking sites to limit the damage of phishing and email spam by giving customers tools to regulate the flow of data.

Yesterday, before New York Attorney General Andrew Cuomo suing Tagged.com story broke, I cold emailed a member of the Gmail team:

Gmail could help web security a lot by providing:
1. Authentication (OAuth) to Gmail address book making it clear that you were not providing your Gmail passsword to a 3rd party web site.
2. Default level of access only provided names and salted hashes of email addresses from address book (possibly 3rd party web site part of salt)
3. Allow only a limited number of actual email addresses to be requested in a time period. I’m guessing ~30 would be a sweet spot.

That would seem to be one possible solution. If this is not a good solution, I think it’s important for your team to look to tackle the problem described below in another way.

EXPLANATION

[Background information described in my "Tagged.com Spam? Phishing? Nice Guys? My Personal Story"]

I’ve seen similar UI at othe web services, where everyone in your address book is selected by default. I think there is an awesome opportunity for your team to create an experience that works well for your partners and protects your customers from the type of mistake described above and more importantly from malicious sites.

Some of the problems that I think Gmail and other online email address book and social networking sites should at least take partial ownership by:

  • Not allowing 3rd party sites to embed login forms. They should use OAuth or a similar approach. (Even on AppEngine — train us well).
  • Having a really clear experience about what data you are giving access to (how pissed your friends might be), and a way to provide only limited data.
  • Providing salted hashes instead of email addresses, so that a person can find their friends on a 3rd party service without having to hand over the actual email addresses of their friends.

I don’t think I read the Google Chrome Operating System announcement until after I sent that email. When I did read the announcement, I thought about how empowering and freeing it will be for our computing to be in the cloud, but I also thought about problems like this one, and how many scary things can happen when you are no longer hold the container(the harddrive in your PC) for your information and data.  There is a lot of design still to be done to create a safe and friendly experience.

SourceForge Projects Moving to WordPress

Open source project hosting has long been a topic that interests me. I stopped recommending venerable SourceForge.net some years ago, as it’s proprietary stack (open source prior to 2001) became crufty and fell behind some of the newer and more agile offerings.

Well, I think it’s time to revisit.

I received an email on Tuesday from the “SourceForge.net Team” with the geeky and lame title of “SourceForge.net feature deprecation upcoming: forums, DocManager, TaskManager, Diary/Notes”. Deprecated should be deprecated from email subject lines, as should negative sounding email subjects.

The meat of the email is the second half (emphasis mine):

We will provide an easy-to-use migration path to move the data to the provided replacements.  We will also provide dumps of this data in case
projects want to do something different with their data.  Additional information on how to obtain or migrate your data will be provided when the
timeline is announced, in a future mailing.

The following applications are due to be deprecated, replaced by high-quality Open Source applications we have in our Hosted Apps offering:

* TaskManager will be replaced by TaskFreak!, dotProject and Trac
(tickets).
* DocManager will be replaced by MediaWiki and Trac (wiki).
* Discussion Forums will be replaced by phpBB.
* Diary and Notes will be replaced by WordPress.

To solicit your feedback on how the migration should be handled, and alternate options you would like us to consider, we are running a survey
for the next 30 days for the user base of each of these applications.  For links to the surveys, please see our Site Status post at:

http://tinyurl.com/q3g8o3

Trac and WordPress (really!) are two of my favorite open source web applications. And all of the applications in that list are highly regarded.

It’s fantastic to see SourceForge getting back to open source — ironic, no. It will be interesting to see how active they are in particulating in those projects. I dont’ think I’ve seen any SorceForge team members participating recently on WordPress’s Bug Tracker (Trac).

It’s also fantastic to see SourceForge engaging their community by posting this on a WordPress powered blog and also using surveys to solicite feedback about the migrations and the tool choices.

I would love to find out how long this change has been in the works and what people made it happen.

I also notice that SourceForge’s own documentation is now in a trac wiki. Wow, as well as Subversion, you can also use either Mercurial or Git.

The SourceForge team has done an incredible amount for open source, and I’m excited to reconsider using SourceForge.net again to future projects.