The guy who hosts my files just called me. He got in a disagreement with the guy who manages his servers, and told me to back everything up because it might not be there tomorrow

Wow. In a split second, all your data can be gone. The forever kind of gone. The problem was – and the reason she called me – was that she had amassed so many files over the years, that it would take days to backup using her internet connection, and she only had hours to get it done. Okay, relax, I told her – I’ve got it taken care of. I can use linux shell commands to download all the files to my server from yours. It will be much faster, and the files will go directly to the new server instead of having to be re-uploaded there, saving some very time-consuming steps.

Okay, I thought. I’ll just login, make a big tarball of all the files and grab that with my server. But her file hosting account did not allow shell access, and probably didn’t have the extra space for an additional tarball of all the files anyway. So I’m stuck with the linux ftp commands – or so I thought. Turns out, the mget ftp command does not recursively download folders on most servers. So the best function to use on a remote linux server that you can’t run shell commands on is wget, because wget also supports the FTP protocol. The usage goes like this:

1

wget -r ftp://user:pass@domain.com

That was going fine, and then the connection was cut-off by the remote server a short way through getting all the files, probably due to some data transfer cap or something. I re-started it, and it cut off again near the same place. So this isn’t enough either, and I still don’t want to do it manually. Luckily, there is a wget flag to ignore already existing files – ‘-nc’. So the whole command to download everything recursively and not re-download files you already have is:

1

wget -nc -r ftp://user:pass@domain.com

Remember to back up often. You never know when you might find yourself in a sudden and unexpected data loss situation, like Ma.Gnolia did Friday, January 30th. There’s a good discussion happening on the SitePoint open thread on data loss that same some good backup ideas and methods, too.

There will be a few additional announcements and activities at the open house, including a speech from Noah Everett, the founder of TwitPic.  My own startup InvoiceMore will have a special announcement as well and a bonus for those who sign-up for OKCCoCo with a full or part-time commitment.  Come out if you can, and don’t forget to register first!  These are exciting times for the Oklahoma technology community!

Basically, the problem is that if you’re using numerically-indexed arrays with a set number that you don’t want to change (like an ID or some other unique identifier), you can’t use array_merge, because it automatically re-indexes all the numeric keys in the array to start with 0 on down in order.  There is no flag or option for the function to NOT do this, but there is another way to achieve the same result using PHP’s little-documented overridden plus operator ‘+’ for appending an array to another array.

So just replace this:

1

$destinationArray = array_merge($array1, $array2);

With this instead:

1

$destinationArray = $array1 + $array2;

Both $array1 and $array2 MUST be arrays or a fatal error will be thrown, so you may want to do some type checking or casting before that line of code. The difference is that instead of merging the arrays together, the second array will simply be appended to the first one with no changes.

Note that the plus operator for arrays ‘+’ is only one-dimensional, and is only suitable for simple arrays. If you need a multi-dimensional or complex solution, Keith Devens has a custom merge function that might work for you.

The dashboard is the place for the most important things to be.  it should be the single place you can view to and get an overview of all the most important things about your website regarding your visitors without having to drill deeper or go through multiple pages or sub-sections.  But the default dashboard Analytics starts you off with is all wrong, and is almost never the information I really want to see.  Let’s see how we can fix this.

The Default Dashboard

Google Analytics Default Dashboard Setup

The first thing to note here is the repetitive information.  The “Visitors Overview” widget is the first thing you want to get rid of.  The exact same information is displayed larger above – no use in repeating the same information twice in the same screen area.  To the right of that is the “Map Overlay” widget.  While this information may be neat to see, it’s not really the real information you need day to day.  That one’s leaving our dashboard too.  So that leaves the two bottom widgets – “Traffic Sources Overview” and “Content Overview”.  These both can stay, but perhaps can be arranged a bit differently.

The Information You Need

For most webmasters, there are essentially two important questions: “How are visitors getting to my website?” and “What are visitors doing on my website?“.  The widgets we select for our dashboard must help us quickly answer those every day questions without having to drill any deeper or take any more of our time.  We also need only the information that isn’t already present on the dashboard that we can’t edit.

A More Useful Dashboard

Google Analytics Customized Dashboard

With a customized dashboard, we can now have the information we really need on a daily basis that help us answer our every day questions:

How are visitors getting to my website?
I can clearly see here that the dominant source is Google searches, followed by direct type-ins and the typo domain that redirects to the real site.  I can also see the top search keywords that visitors used to reach my website, which is very useful for measuring SEO effectiveness.

Another dashboard widget I like to have to help answer this question is the “Referring Sites” widget.  This tells me the actual websites referring visitors to me through links.  This is especially helpful if the site gets mentioned in a blog post or article so you can know about it quickly.

If you run any Google Adwords campaigns, there is also an Adwords section and a dashboard widget for campaign clickthroughs and conversions that may be helpful.

What are visitors doing on my website?
This question may take a bit more analysis to answer because there are many parts to this.  The “Content Overview” widget helps us spot the most popular pages and areas of the website, which helps us solve most of this question.

Taking the analysis further, you may want to setup conversion goals and add a dashboard widget to display the number of daily goal conversions.  You can also add widgets for the entrance and exit pages, among many other things.

Conclusion

Most of the time for most of my websites, I have only the widgets above shown on the customized dashboard image above.  The great thing about Google Analytics is that it can be customized with any widgets you want to display.  So although the default widgets are not particularly useful on a day to day basis, they can easily be replaced with widgets containing information that is useful.

One more final thing about Google Anlytics is that the date range can be customized as well.  In the screenshots above, stats for the whole month are displayed.  If you have a larger website, you may want to customize the date range to a week or even a single day to get a better understanding of what’s going on most recently.  The drawback of course, as anyone that uses Google Analytics knows, is that the stats aren’t live or real-time.  There is a 24-hour delay because all the stats are compiled and processed at the end of each day.  I’m not going to complain about it because the product is free, but just remember that if you do choose to view a single day, you will have to select the previous day instead of the current one.

<button type=”submit”> vs <input type=”submit”>

You would think both elements would behave exactly the same, because they’re both basically the same thing – the “submit” button on HTML forms.  Turns out that in Internet Explorer 6, they don’t behave the same at all.  Surprise, surprise.  The issue has to do with using multiple submit buttons on a form which can perform different actions:

1
2
3
4
5

<form action="somewhere.php" method="post">
<p>[Displayed cart contents with editable quantity fields]</p>
<button type="submit" name="update">Update</button>
<button type="submit" name="checkout">Checkout &gt;</button>
</form>

The standard behavior of forms is to only pass the submit button field that was clicked by the user. So if the user updated some quantities and clicked on “Update”, only that <button type=”submit”> would be posted back to the server, and the “checkout” field wouldn’t even be set. So the code on the server to check which button was clicked might look something like this on the server:

1
2
3
4
5
6

// Which button was clicked?
if(isset($_POST['update'])) {
    // Update cart quantities
} else { 
    // Save cart and redirect to checkout page
}

What’s Going On?! @#%$ IE!

Turns out, IE6 actually passes ALL <button type=”submit”> elements back to the server on postback instead of just the active/clicked one. As a result, the “update” element will ALWAYS be set for ie6 users, and the “else” condition will NEVER be executed. They won’t be able to make it past the cart page, and won’t be able to buy your products or continue to the next step.

The Solution

So the solution here is to change all the <button type=”submit”> elements to <input type=”submit”> elements instead, because those are actually handled correctly in IE6.

1
2
3
4
5

<form action="somewhere.php" method="post">
<p>[Displayed cart contents with editable quantity fields]</p>
<input type="submit" name="update" value="Update" />
<input type="submit" name="checkout" value="Checkout &gt;" />
</form>

So much for the flexibility of being able to use HTML markup inside <button> elements.

OpenBeta is an un-conference — a participant-driven community event aimed at exposing ideas and creating opportunities by connecting smart, passionate and motivated people with different expertises.

The event will be held on Thursday, October 30, 2008 at the IAO Gallery in downtown Oklahoma City at 7:00pm.  It’s going to be a night of good fun and networking with like-minded technology-oriented individuals in Oklahoma.  I’ll be giving a quick 5-minute lightning talk presentation myself on my upcoming InvoiceMore project and what makes it different from competing online billing and involving services that already exist.

Please consider attending, and have those networking skills ready!

Procedural to Object-Oriented: The Benefits of Using Object-Oriented PHP

Learn the power of object-oriented programming in PHP5 and the many benefits it offers over the more traditional PHP procedural programming style.  This session will include a light introduction to object-oriented concepts and will provide real-world concrete examples of the benefits it can offer you and the PHP projects you work on.

I will be speaking on October 9th at 2:30pm, and the presentation will last for roughly 75 minutes (60 minutes to speak, and 15 minutes for Q&A).  That’s a good chunk of time to fill, but there’s a lot on this topic that will need to be covered.  If you’re thinking about getting into object-oriented PHP programing or would like to learn more about it, please attend.  I will try my best to make sure there is at least something that everyone can learn.

You can also view my page on the conference website to read a short biography and get more information on the event.  Hope to see at least a few friendly faces there!

P.S. – I plan on posting my presentation slides on this website after the event just in case anyone missed anything important or was unable to attend.

UPDATE: The conference is over, and I have posted the powerpoint slides in another post for those that are interested in the presentations I made.

The Minutes Pass Quickly

All of the Woot-off checkers on the market do the same exact thing: check the current deal at woot.com, show all the relevant information on that item, and tell you what percentage of the item is sold so you can anticipate the next item.  The problem is, some of those items are really great deals, and some of them aren’t.  You never really know just how great a deal the item is until you do some research and price comparisons on it.  But that research takes a few minutes of time.  And while that may not sound like a lot of time, items during a woot-off occasionally sell out in less than a minute.

Differentiation: Inline Price Comparison Results

The goal for a great Woot! checker should be not only to simply regurgitate information directly from the Woot! data feed, but to also provide the information necessary for users to make a purchasing decision quickly.  After all, the very reason for the existence of any Woot! checker is so you don’t miss out on a great deal.  Why not take that idea to the next level to give your users at least a 15-second advantage by eliminating steps in the research process? That’s where Woot! Compare comes in.

Woot! Compare fetches the normal information from the Woot! data feed, but then it also goes a step further and fetches shopping comparison results automatically based on the Woot! item name.  These shopping comparison results are displayed inline right next to the sale item so you can compare prices to ensure you’re getting a great deal.  Since the shopping comparison process is completely automated and sometimes the product name doesn’t have a specific model number, the comparison results can’t be guaranteed to be an exact match to the sale item all the time.  That’s why the product images are so critically important in this situation.  So even if the shopping comparison search does happen to include an unrelated or non-matching item, the user can quickly dismiss it and still make a good evaluation of the Woot! item’s worth based on the other product results displayed.

So now hopefully you will be able to spot that perfect holiday gift and be able to make a purchasing decision before the opportunity passes you by and the item stocks out.

Happy shopping!

Resources

• Woot! Compare – The end result of this experiment
• Woot! RSS Data Feed – Clean XML data for the sale item
• Google Base – Especially Google Base API
  
• Shopzilla
• Shopping.com

The primary focus of this blog will be on my daily adventures in advanced PHP coding, Javascript, and general web development.  The goal will be to supply code, resources, and helpful articles to those who are also coding or learning object-oriented PHP, Javascript, HTML, CSS, or other web technologies.  And of course, there may be occasional deviance from the main focus, though my posts will still be web-related and relevant to anyone in the Internet industry. My adventures in blogging begin… Now.