For months I was annoyed at the inability of my WordPress installation to automatically upgrade both plugins as well as the core installation. Every time I had to do a core upgrade of WordPress, I would manually download the zip file, unpack, copy the wp-contents and wp-config files and then take care of any outdated plugins. I tried many, many times to open the FTP ports on my instance and do the upgrades, but time after time I got the same annoying connectivity error.

I decided today to end all that, as I am getting ready to get back to blogging (after a couple of months hiatus, related to a project I’ll share with you later) and wanted to start with the new WordPress 3.0.

First of all, I needed to figure out how to open the FTP server of my instance, since I don’t have it open all the time. I highly recommend to everyone to keep FTP closed, and only open as needed (minutes is better) as FTP can be easily be exploited by hackers.

FTP works with multiple ports. On one hand you have the standard ports, which are port 20 and 21, and on the other a “passive” control port that is randomly picked per session.

I recommend that you use vsftpd, which is standard on most Linux installations. You can start and stop vsftpd using the command: sudo /sbin/service vstfpd start (or stop).

First, you will need to edit your vsftpd config file by editing /etc/vsftpd/vsftpd.conf. At the end of the file, insert the following snippet:

# Passive support
pam_service_name=vsftpd
userlist_enable=YES
tcp_wrappers=YES

pasv_min_port=51000
pasv_max_port=51999
pasv_address=201.201.144.144
pasv_addr_resolve=YES
cmds_allowed=USER, PASS, QUIT, PASV, RETR, SYST, PWD, NOOP, CWD, STOR, LIST, DELE, MKD


The pasv_min_port and pasv_max_port can be any range you pick, as long as they are available on your instance. Then, on the pasv_address put your instance’s Elastic IP address or publicly visible IP.

Next go to the AWS Management Console, and under Networking & Security select the Security Groups option and make sure to add two entries on the list of ports: one for TCP ports 20 to 21 and another for TCP ports 51000 to 51999. See the graphic below:

Now, make sure you start vsftpd (using sudo /sbin/service vstfpd start) and check from an outside computer that you can log in to your machine. You need to make sure that you have a user that will be chmoded to your root wordpress installation (using the command useradd, as described here).

One last thing on the FTP side is that the apache user has access to your whole installation (as the upgrade process will be driven by the WordPress program which runs as apache). You can do that by issuing sudo chown -R apache:apache blog/.

If you try doing the upgrade now, you could still receive this cryptic error message: “Unable to locate WordPress root directory”. I found the fix, via this blog, where you basically add the following snippet at the end of your wp-config.php file:

if(is_admin()){
add_filter('filesystem_method', create_function('$a', 'return "direct";'));
define('FS_CHMOD_DIR', 0751 );
}

Try now — if the WordPress Gods are smiling at you, you should be able to upgrade via FTP. A reconfiguration of the WordPress may be necessary, so make sure to do a backup of your blog DB. But I’ve got to say, once I saw WordPress’s 3.0 Dashboard only after a few seconds, I was glad I spent the time to finally get this to work. I also upgraded a couple of plugins without any problems.

After you’re done, remember to shut down vsftpd — don’t let it running on your instance to avoid possible hacking, by doing sudo /sbin/service/vsftpd stop

Let me know if this worked for you. I’m looking forward to a whole lot of blogging in the following months!

There is a natural cycle in Social Media applications, where there’s an initial excitement (the romantic phase), a leveling of activity (the wedding phase) and hopefully the “till death do us part” phase, where the application becomes part of our lives.

But most often than not, there’s a divorce phase. The application just doesn’t measure up to our needs. Or the community on it changes.

The latter is what’s happened with FriendFeed. Even though the service is up and running (albeit with some ongoing server hiccups), this brilliant post from Akiva Moskovitz says it all: the FriendFeed community has stagnated.

I was one of the users who fled the community as soon as I heard that Facebook had acquired FriendFeed and tried, in vain, to move my “social furniture” to a new home. I tried Facebook, I tried Twitter, I tried Pip.io.

I was even very bullish about Google Buzz. But after really trying it for a number of weeks, the truth is Google Buzz doesn’t work as an aggregator, but it’s rather a commenting platform for original content. The problem is that its algorithm seems to favor people with lots of followers, but for users with a few hundreds of friends like me, it’s sort of a dead town. I get more value reading my feeds in Google Reader than going to Buzz and reading about the same things, or read what DeWitt Clinton, Louis Gray or Thomas Hawk are talking about.

The truth is that none of these places felt like home. We really got spoiled during the golden time of FriendFeed, when you posted anything, and you immediately got feedback, amazing comments and different points of view.

Today, FriendFeed activity is still strong with a few users. But, going back to Akiva, it’s the same small group of people commenting and regurgitating their points of view. I am close with a lot of them and care about them. But the truth is that FriendFeed is not that active place anymore. Even Louis Gray, one of FriendFeed’s last faithful defenders, talks about the very apparent decrease of traffic coming from the site.

At the end of the day, I still think it’s Facebook’s game for the taking. I am sure the FriendFeed team is working hard in turning Facebook into the FriendFeed for the masses. Facebook already surpassed Google in U.S. traffic, and there is a community there (including a lot of my FriendFeed peeps). The main obstacle is that Facebook’s present UI plainly sucks to do effective sharing of content.

Either that, or a completely unknown startup sweeps everyone’s feet. It’s happened before and it’s going to happen again.

Until then, I will be sitting here in the middle of the huge social crater called FriendFeed. It’s still the place I call my social home.

Photo from NASA – Jet Propulsion Labs

With Google Buzz, users have found their inboxes converted instantaneously into a social hub. Google’s bold move has turned the web into a whirlwind of blog posts appraising or criticizing the service.

I will leave that part out of this post, as I feel it’s not really fair to evaluate a service that has two days of life, but rather will concentrate on some of the ways you can maximize the use of Buzz if you want to give it a try.

Hide your following/followers

So I agree it is not a good thing to show to the world the contact info and email (Google profiles are generated from your Gmail account). Because Buzz is not centered on follower numbers (like Twitter is) you should definitely and first of all hide that information.

The way to do that is by editing your Google profile and checking off the option “Display the list of people I’m following and people following me”.

That way anyone visiting your profile won’t be able to see your contact cloud and you can be safe they won’t get spammed because of you.

Keyboard shortcuts are your friends

Using Buzz with your mouse (or touchpad) can be a pain, as posts can be short or long based on the number of comments they have and embedded media.

If you are a Google Reader power user, you already know about the keyboard shortcuts there. You’ll be happy to know that they too work on Buzz.

First, you need to enable keyboard shortcuts on Gmail. Just go to the Gmail settings page and enable keyboard shortcuts.

These are the two keyboard shortcuts I use the most: “n” and “p” allow you to move to the next and previous buzzes. You will see a little arrow on the left top side of the buzz to let you know which one is selected. That way you can navigate through the buzzes fairly quickly.

The second one is the “m” key, which mutes a buzz. That way, if you are not interested in the comments or activity of a buzz, you just mute it and it will be forever archived, unclogging the good stuff you do want to follow. I also believe by muting buzzes, Google’s system can start to learn what’s interesting to you and what isn’t (although that’s more of a hunch than a statement, maybe someone from Google can confirm that).

Don’t import echoes

One thing that gets really annoying is echoes of content from the same source. For example, if you are importing your Google Reader items to Twitter and you are also importing Twitter to Google Buzz, users will see your shared Google Reader items twice. Same goes for FriendFeed (it’s usually a safe bet to not import FriendFeed into Buzz, unless you’re just posting original content there). Just make sure you’re bringing the good stuff just once to your buzz community.

Post from email

A really cool feature that Buzz offers from the get-go is email posting. Just email to buzz@gmail.com from your Gmail account (either web or mobile) and the Subject of your email becomes the title of a new buzz. You can also attach a picture file and it will be added to your buzz.

You can also fine tune who sees those email posts, by going to your “Connected Services” page (just make sure you’ve posted at least one buzz email) and select which group sees those.

Filters are your friends

There are some cool things you can do with filters, both in Gmail as well as in the Buzz search areas.

As indicated by Lifehacker’s Adam Pash, you can have buzzes skip your inbox, so that you’re not distracted by those notification emails.

However, I changed it a bit, because I wanted two labels: one with the buzzes I had participated in and another where I had activity of buzzes I started.

For the first one, I did created a “Buzzes” label, doing a “-From:Jorge Escobar” (notice the minus sign on the front of the name) on the To: and “label:buzz”.

For the second one, I created a label called “My Buzzes”, where I put the “From:Jorge Escobar” and the “label:buzz”. This mimics a little bit the FriendFeed “My Discussions” page.

Another cool tip I learned from my friend Stefan Svartling is the filtering by services. For example, on your Buzz Search input, you can type “jungleg -twitter” and it will show all of my buzzes excluding the Twitter posts. You can filter other services like flickr, picasa, etc. Maybe someone can post a link to what all the services names are.

I am excited of what comes down the road with Google Buzz and I’d love to hear other tips you find out as you discover the service.

Remember to connect with me using my Google Profile.

Happy Buzzing!

Yesterday I saw the realtime video feed (thanks to @scobleizer) of the unveiling of Google’s new phone, the Nexus One. The phone looks great, with some really cool features and confirming that Google is really lining all its guns towards the mobile space.

But the announcement wasn’t revolutionary in any level. It’s a great new phone that’s powered by any cellular network. You can buy it straight from Google with no string attached. Ho-hum.

Before the announcement, my friend @joeles thought that the phone would finally break the dependence on mobile voice plans, working off of data plans exclusively using Google Voice (@mona was talking about the same thing on this blog post and got really good responses on FriendFeed).

I believe on the long run this is still the case, but it won’t be running on top of Wifi or data plans. I think Google will be utilizing the “whitespace” frequencies that were freed from TV signals to enable data communication on the Nexus or any upcoming Google phones.

The big question mark as usual is the business relationships that could potentially be broken by this shift. T-Mobile could have enforced some sort of contract with Google to stop this from happening. I don’t think that would be Google’s way of doing things, but you can never know. The fact that Google is selling the phone directly is a good indication that they are not going to be tied up with any carrier.

The mobile disruption did not happen just yet. But knowing Google, you can bet it’s on its way.

Today I noticed Compete.com had come out with November stats. My blog had a spectacular month (and if you read it, you know why), but my intentions were other: to see how FriendFeed fared in a key month after its Facebook announcement has died down.

The graph doesn’t look good for FriendFeed, which has lost an additional 20% of its audience.

The question we, as FriendFeed fanatics ask ourselves is why? Why is a product that is so unique been left for dead all of a sudden.

One camp would point to the Facebook announcement. Who wants to spend time and energy on a satellite service, knowing that the resources are aligned with the parent company. This is interesting because it would tell a lot about user’s behavior in this Real Time Web: they want a product that evolves, even though the product is perfectly fine.

It’s the same feeling I get with the netbooks, the crunchpads and the Apple iTablet’s: we’re always looking for better, faster, stronger.

But why?

If the product or site or service does what it advertises, why does it have to keep development on a frantic race to over-development?

But then you see Twitter. They haven’t changed that much since they started. Yes, there are a lot of applications that leverage it, so it has evolved to become more a platform than a service. Still they are not losing the audience that FriendFeed has.

Clive Thompson wrote a great article on this month’s Wired about how startups these days are following a safe cycle. They launch something quick, they aim to build an audience, they sell to one of the big boys and then die. Rinse and repeat.

The problem is that none of this startups are aiming to change the world. They are following the iPhone App paradigm. Small, sweet and sold under two bucks.

Did FriendFeed’s intention all along follow this recipe? Not in my opinion. The service was really ingenious and they developed the best in class search, bookmarklet, and other features that I haven’t seen developed since.

Maybe FriendFeed was too much of a destination. They did have an API but developers didn’t flock to do stuff with it. Marketers didn’t ask for datamining. It’s so strange.

At the end of the day, we still don’t know what is the certain future for our beloved application.

But we’ll be here until it dies or until something better comes along.

I’m betting we won’t see the latter.

There is a rush when you see a post that gets a lot of attention. I remember Hutch Carpenter experiencing something similar when he had his “Cisco Fatty” post.

But after the moment passes and you look back, there’s a bunch of negative things that slowly emerge from the experience. I will share with you some of the ones I have thought of.

But first, a little backstory. On November 17th I received an invitation to try a Microsoft installer, which I mistakingly thought was related to Azure, Microsoft’s Cloud Computing service to come out next year. I wrote a post about how this could be a great thing for Cloud providers, allowing non-technical users to deploy cloud-aware apps using an iPhone application metaphor. A fellow FriendFeeder posted it on Hacker News, and from there, the post registered an amazing flurry of visits. The post stayed on the home page of Hacker News for most of the day.

So here’s what I gathered as the post’s traffic died down 3 days later.

• Spikes on Analytics distort your overall view. I am now left with a small horizontal line that doesn’t let me really get a feel of how the blog is doing. I wish there was an option to take that day out of my report, so that I can see the real picture.
• Most of the traffic doesn’t stick. The before and after subscriber numbers didn’t change at all. These types of traffic are mostly “one visit and forget about your blog” type. The subscriber number did spike as well, but I think this is a false inflation that Feedburner detects, but it’s not real.
• It gives you a false sense of success. Suddenly you see that day’s traffic and multiply it by 30 and you’re now up there with the big boys. It doesn’t happen like that. You need to keep writing, for years, before that happens my friends. You also don’t feel like blogging for a while, because your monthly metric is now so much higher than your previous month’s — even though the event happened in the middle of the month.
• You get very tired for days after the event. I stopped thinking about new posts for days because my mind was still reeling from the false success factor.
• You suddenly are challenged to do an even better follow post, which is unlikely. I guess this is related to the previous one, but you are now defeated in your own turf, pressuring yourself to follow up with an even bigger post.

At the end of the day, there’s little benefits from these spikes, and you shouldn’t put too much energy on them. You might have a nice graph to show yourself on the monthly reports, but the real takeaway is that these are isolated events that should be mostly ignored and forgotten by the next day.

The only positive thing you should learn from a spike is trying to figure out if there’s a formula you can extract from it that you can apply for future posts. In my case, I was extrapolating a “now” event into the future, which might have captivated user’s imagination, which is something I often try on my blog. It would look like this is something I’ve got to keep trying.

Looking forward to your comments.

As I read report after report on Google’s Chrome OS, I have to say I feel a little let down. But most importantly, I’m hesitant to call it a real Operating System.

According to the official post:

[Google Chrome OS is] an open source operating system for people who spend most of their time on the web.

Don’t we all spend most of our time nowadays on the web? What is not to love?

The problem is the rest of the time when we are not on the web.

You see, Chrome OS doesn’t write anything locally, it leverages the cloud for all write operations.

I haven’t read the specs closely, but I assume ChromeOS will have Gears enabled, so that when I’m on the subway, I can still work on my Google Docs.

But there’s been more than one time when syncing failed and I was able to copy and paste the contents of the document in a notepad file and save it locally.

Then there’s the fact that you can’t run any applications that require hard drive installations. If you see my desktop right now, Chrome is handling most of the stuff I do on a daily basis (Gmail, Twitter, Google Reader, Blogs, FriendFeed, WordPress) but if you look at my taskbar, you’ll see Outlook (and I need it, because the web-based Outlook sucks), and I am running XAMPP, because I develop web applications locally, specially without needing to have internet connectivity (again, the subway is my second office). I also use Fireworks often, which would have to be replaced with a cloud-based image editing system.

And this is all running on an MSI Netbook which is never turned off.

My point is that it’s awesome that we can do more and more on the cloud. But there are situations when we’re going to need to work locally, and installing an OS that has been striped of that functionality is like tearing the back seats of my car just because it makes it lighter. At some point I’m going to need those seats.

See the video below to get a better idea of what ChromeOS is in a nutshell.

Additional articles: Was Chrome OS a Disappointment? – Why Chrome OS will fail — big time

I just received an email invitation to try out a new application. I get a lot of those these days, but this one was different.

It was from Microsoft.

It piqued my interest. A Web Platform Installer? Microsoft doing PHP?

I went to the URL provided and I was blown away with the concept behind this application.

Basically Windows could potentially introduce point-and-click cloud computing for the masses in a way that resembles the iPhone application directory but for web applications, once their Azure cloud service is launched.

According to my friend Roberto Bonini, Azure is just the beginning of a frontal attack for the Cloud Computing crown:

The way I see this evolving is that basically you could launch a cloud-enabled version of the Web Platform Installer add an application from the gallery and launch it on Microsoft’s Azure Cloud and that’s it. The application basically handles the database, frontend, and serving in a matter of minutes. Of course developers would need to modify their payloads to be cloud-aware, but this is not something crazy.

There is already an application gallery that you can see. Matt Mullenweg was quoted today that WordPress is one of these applications, so bloggers can start their self-hosted blog in minutes, and there are many more already listed, like SugarCRM and mojoPortal. Microsoft is inviting developers to submit their applications to potentially be run by millions of users.

Windows Azure won’t launch until January 1st, but Microsoft is working hard to position itself as the defacto provider of cloud computing for the masses. This is going to be the Microsoft of the future. You can tell they know this is their new business model: cloud applications.

Microsoft has found their new Outlook.

Author’s note: I have edited the article to express my ideas better in terms of where I am extrapolating and where this is an existing application. For the record, I am big fan of Open Source stack, and as a matter of fact this blog runs on EC2, and I have million hits applications currently running for commercial enterprises on the AWS environment.

In talking to different startups in the past weeks, it’s very clear to me that businesses haven’t grasped yet how the Internet has shifted from the destination paradigm to the platform paradigm.

In a post titled “The Web in Danger“, Anil Dash compiles and adds to the thoughts of Tim O’Reilly, Doc Searls and Chris Messina about how the web is in danger of losing its essence: the destination URL.

So far people have thought of websites by the URLs they enter on their browsers to consume its services. But today, they are thinking of businesses as omnipresent services. They want to be able to do everything they  normally do on the URL, using their iPhone or on Facebook or on their Chrome OS powered netbook.

They want to fire up your application using an icon; not enter an address on a URL window.

I don’t want to paint this post with undertones of doom: I think we will still have browsers and URLs and it’s not like the Web is going to disappear.

But, thinking as a business, these are some points to think about for your technology roadmap:

1. How pervasive is your business? Can people interact with your business only on its URL? I would recommend at least doing a mobile-compatible version of your site with a minimal set of tools available. I would also think about doing a Facebook application in the short term (although I’d wait until Q1 next year after Facebook decides what it’s going to do with their new apps roadmap).
2. What is your user authentication process? If you are still requiring users to register for an account (and on top of that, you ask for lots of information) you are very much inciting prospective users to leave. You need to offer at least Facebook Connect. If you also offer Twitter and/or OpenID, even better (hey, there’s RPX, so no excuses).
3. Do you have an API in place or in your roadmap? The more people you get to use your tools, the better. I agree that some businesses don’t need an API, but I see lots of business that should have an API or a more robust API than what they have.

In the end, I think there’s no danger of losing the Web’s openness. It’s true that some companies are blowing up in terms of growth and look like monopolies in the short term, but if you think about all of the big ones (Twitter, Facebook, Google) their real strengths have been to become platforms and not portals (remember AOL? I don’t either). It’s up to developers to continue creating ecosystems on top of these platforms, but if the past is prologue, I think there are still many opportunities for new platforms to come and transform, yet again, our good old Internet.

I just spent the last ten days trying to do an import on Amazon S3 using their Import/Export service. Basically Import/Export allows you to send a drive to Amazon via snail mail and they will hook the drive to their system and import the data locally. It’s much faster than trying to upload the first snapshot of the data via the good old internet.

The problem is that the service requires you to produce a manifest file (to let them know what bucket to load unto as well as some other information) and a signature file on the unit itself that verifies that you are the legitimate owner of the process.

The official documentation can be read here, but I will give you the “real” how to.

The first thing you do is to make a manifest file. Instead of writing the file, I opted to use the S3Fox Firefox extension, which writes the file for you. This is a problem, and I will explain to you why later on.

You will be better off writing the file yourself, and it’s not that complicated. Just write the following:

I added a couple of extra variables.

Because I wanted the files to go into a specific path, I added the “prefix: myownpath” so that data is sucked in there. The “deviceid” is the serial number of the unit you’re shipping.

I also listed the types of files (which is kind of weird, but Amazon requires it). Here’s the list I did:

contentTypes:
flv:application/x-shockwave-flash
mov:video/quicktime
wmv:video/x-msvideo
mp4:video/mp4
pdf:application/pdf
ai:application/postscript
zip:application/zip
mpg:video/mpeg
mp3:audio/mpeg
psd:image/psd
png:image/png
jpg:image/jpeg
gif:image/gif

Now here’s my nutshell of advice to y’all: Make sure you create this file on a Linux/Unix environment and not using Windows programs or S3Fox. Again you’ll see why later. Create this file using vim or emacs and then save it as “manifest.txt”.

Send an email to “awsimportexport@amazon.com” with the subject “CREATE JOB” (all caps) and attach the manifest.txt file to the email.

Amazon will take its time and then reply back with something like this:

SUCCESS
JOBID: J5N9M
deviceId:97333A9360CC3

You will now need to create a SIGNATURE file. Download the signature generating tool from Amazon and put it on /tmp or create a folder and unzip it.

Go to the directory and run the following command:

java -jar lib/AWSImportClientSignatureTool-1.0.jar manifest.txt J5N9M SIGNATURE

As you can see, you need the JOBID that Amazon sent on the email. The “SIGNATURE” is the name of the file. Leave it like that, all caps.

The script will ask you for your Amazon AWS Secret file. Copy and paste it and then it will generate the SIGNATURE file. Copy that to the root folder of the unit you will ship.

Now you just need to print out a shipping label to put inside the box with the storage unit, and UPS/Fedex to Amazon. Here’s the documentation for that.

You should get an email when Amazon gets your unit and starts the importing.

So for the HOW-NOT-TO-DO-IT guide: I created the manifest file using the S3Fox tool and saved it on my Windows Desktop. I opened it using Notepad to double check everything was fine.

I sent the CREATE JOB email and received the JOBID. I then used S3Fox again to create the SIGNATURE and shipped the unit.

When I got the email from Amazon that they had received the device, it had a failure: The SIGNATURE file was invalid.

I had a back and forth with AWS for 5 days. They were really helpful (Thanks Chesley and Eric!) but I wasn’t clear what I was doing wrong. It turns out that when the manifest file is generated and saved on Windows (or opened in Notepad) Windows will insert character returns at the end of each line, and the SIGNATURE generated in S3Fox takes this into consideration. But when Amazon processes the manifest, it strips out these character returns and generates a SIGNATURE that, when compared to the one you send, is not the same, and thus marks as invalid.

So make sure to just do the whole process on your Linux terminal and not on your Windows GUI and you’ll have your backup done in no time.