After a bit of digging, I found the location where Chrome stored user data:

• On XP – C:\Documents and Settings\<User Name>\Local Settings\Application Data\Google\Chrome\User Data
• On Vista - C:\Users\<User Name>\AppData\Local\Google\Chrome\User Data

The User Data folder contains three files: Local State, Safe Browsing and Safe Browsing Filter, along with a folder called Default. Default in turn contains your browser cache, plugin data, and all of your cookies and history data. To move my profile over to my new computer, I copied all of the files and folders under User Data on my XP machine, and moved them into the User Data on my new Vista machine (all of the files were nearly 100mb after only four days of use, which will give you some kind of idea about the amount of indexing going on in the background). When I next started Chrome on my Vista machine, it was identical to the app on my XP machine, down to most popular sites, history and cookies. I even started writing this post on my XP machine, and then continued it on my Vista machine without having to log in again into my WordPress admin.

In the end this was pretty easy to do. Though the ease of profile transfer could in turn make it easy for someone to steal someone else’s identity – after all, the cookies file (presumably a sqllite db or something similar) was only 256KB, and merely dropping it in the new User Data allowed a complete transfer of identity (perhaps a good security feature would be to allow the \User Data\Default\Cookies file to work only on the originally installed instance).

Likes

• Fast. Very fast, and small memory footprint.
• Each tab is a different process. This will make it very easy to any single tab that is using lots of memory, without having to close the browser (unlike in FF).
• Tab positioning over the address bar (as oppossed to underneath in FF) seems more natural.
• No header or footer bars. Do we really need to waste vertical screen space just to tell me the name of the program and reserve space for status messages? Nope. Here there is no header bar (functions like close/minimize are squeezed to the right side of the tab area), and status messages and urls in links appear in a temporary popup box fixed to the bottom left corner of the browser window, when necessary. Seems like a very good use of screen space.
• Address bar has the domain name of the site appear in regular type, with the rest of the url appearing in a lighter type. It highlights the domain, which very quickly highlights where you are.
• Real warnings for potentially problematic domains. In IE and FF, you just get a small red box in your address bar. In Chrome, when you go to a page that is potentially troublesome (example: loads resources from a domain associated with malware) you have to go through a confirmation screen before loading the site. Seems like a much better way of implementing this than the halfway solution in IE and FF that everyone will tend to ignore.
• Useful built-in developer extensions (View Source, Debug Javascript, Javascript Console, Task Manager).
• In-page search is slick and improves on FF’s implementation. You now see all occurences of the search term highlighted on the page at once, while maintaining the ability to enumerate through the bunch.

Dislikes

• Non-existent bookmarking. I like my delicious add-on and bookmarks toolbar in Firefox as well as my different bookmarklets (Note in Google Reader, delicious, Seed Newsvine). It gives me easy access to the pages that I use frequently (I know that they are stored in the Chrome history, but often just clicking on my bookmark link is more efficient than going to remember the title or url, typing it into the address bar and sorting through the results to find the one that I want). I take it back. Ctrl-B attaches a bookmark bar to the bottom of the address bar. 
• No page titles. Since there is no header bar, the page title is squeezed into the tab for that browser window, which in almost all cases is too small to see the page title. 
• Clicking my mouse scroll bar doesn’t bring up the scroll pointer like it does in almost every other application.
• I like having the search box separated from the address bar like it is in FF. 
• No FF add-ons. No Firebug, Greasemonkey, Adblock, FireFTP or IE Tab).
• No built-in support for RSS. I would at least have expected them to include an auto-subscribe to Google Reader.

This is what I can come up with after using Chrome for 3 hours. So I like it better than IE7 (haven’t tried 8 beta yet), though it does not beat out FF3 (yet).

Some of the features that attracted me are the ability to use one interface to publish on a number of different blogs at the same time, through a desktop client, without having to log in and out of different admin sites (I currently write or contribute to 4-5 different WordPress blogs with varying degrees of frequency, and the ability to post from one place makes it much easier to add content to different sites).

I was also curious to see how well Microsoft has gotten one of their products to interact with software of non-MS origin (and with PHP open source systems like WordPress). Compared with others, Microsoft very often seems to have a hard time opening up their products to others (just look at the new Live email vs. Gmail: Gmail lets you forward your email anywhere, and retrieve your email through any program using POP or IMAP, while Live email only lets you forward to an email address that is part of another Microsoft domain, like msn.com, hotmail.com, but other than that gives no options to extracting your email automatically outside of the Microsoft servers).

Here are some things worth noting so far:

Download and Setup: Easy and painless. Install went quickly. I entered the url for my homepage, my username and password. The software automatically determined that I was using WordPress, downloaded theme information and set itself up. From when I clicked download, I was ready to write to my blog using Live Writer in less than 10 minutes.

Web Preview: A very handy feature, allowing you to preview your post within your current blog theme. It actually did a very decent job of rendering the post the way it should (judging by Firefox). (This is in Web Preview mode, giving you a read-only snapshot of how your page would look. The Web Layout view mode however did not work so well for me. From what I gather, this is supposed to let you compose your post inside your blog theme. In my case, it put the title of my post inside the header, making it impossible to see what I was typing.

The Basic Stuff: Hyperlinks, pictures, trackbacks, basic editing functions – all of it works well, and is pretty easy to figure out. Though this is not something that will make a product shine in the market (even spell check is common nowadays, both in Firefox and WordPress), lack of these features (or a bad implementation) will kill a product from the get-go. (One WordPress caveat: although WLW includes native support for different tagging systems, it does not yet allow you to post tags to WordPress 2.3+. There are workarounds for this.)

Plugins: Although WordPress features a very powerful plugin system, it is most often used to improve the output of your blog, not the UI for entering posts. WLW plugins on the other hand are all centered on helping you get different types of data easily into your posts. I see this (as well as the automatic connectivity to different blogging systems and accounts) as the feature that will make this product really stand out. So far I have installed two different plugins (automatically replacing text formatting methods that I had previously had to code into my template, css and html):

Insert Code:

public static string HelloWorld() {
  string s = "Hello World";
  return s.ToUpper();
}

(Note: I first tried using the Insert Source Code Snippet plugin, but this was buggy and added a whole bunch of superfluous brackets. Afterwards I looked some more and found the Insert Code plugin by shahineo – easy to use and as you can see, produces some nice output).

Insert LTR Text:

זה יעזור לי הרבה לכתוב בעברית בבלוג שלי

All things considered, I am pretty satisfied with my first experience using Windows Live Writer. If you are reading this, it means that the post was also successfully posted to my blog through the interface (yay!). I think that I will be making it my primary blogging platform over the next couple of months. Now, if only I could use this to post to Newsvine…

Ben Rockwood has posted some speculation as to how (mt) may have set up their whole new service offering:

The Grid magic is this: store all use data on NFS so that no matter which system you connect to you can access the data. Then spread your vhost configuration to all hosts in the “grid”, so that any system can serve your data. This system is therefore highly scalable because adding an additional node to the “grid” is trivial and reliable because if one system dies, big deal. But this means that you require two things to make it work: really good load balancers and really good NFS storage. And by good I mean very reliable and extremely fast.

The NFS in this case would be something from BlueArc. Although some people (in the comments and elsewhere) have expressed concern as to the viability of this setup, that doesn’t concern me too much. I am not running anything really mission-critical on the (gs) servers (and even if I was, I do not see it as being any less vulnerable than the traditional shared hosting setup where your site is on one physical box with 2000 other sites).

It does make for an interesting read though.

(Update: right after I wrote this, MediaTemple experienced some more problems with the (gs) service. You can read about what happened and the solution here. Everything seems to be working fine on my end now.)

For me, it came down to MediaTemple and DreamHost. I have used DreamHost for a client site in the past, and found that they are a company that is very concientous of their customer’s well-being. They go out of their way to give easy access to different server features on their (somewhat bizzare looking but very useful) control panel, and overall are very open about different issues that are occuring with their servers (and are not afraid to take responsibility when things get messed up and it is their fault). Overall, I have been pretty impressed with their hosting, as well as customer service. (And am reassured that if problems occurred on a server, that they would take care of it ASAP, unlike my current webhosts).

Then there is MediaTemple. I first found out about them after seeing the little (mt) links on various popular blogs. Whenever people refer to a company that provides them paid services with little emblems, proudly giving free advertising, it says something. Since then, I have read more and more about this hosting service, their top-notch customer service and quality hosting.

The thing that clinched the decision in the end is MediaTemple’s new GridServer (gs) hosting package. This is a revolutionary hosting setup (which everyone will have to do in a few years in order to stay in business). Your sites are no longer tied down to specific hardware, that comes with its own limitations. Instead your sites are hosted by clusters of servers, that can scale instantly to handle large spikes in traffic (ie: if you are dugg, slashdotted and redd at the same time, your shared hosting plan might not be able to handle it, while (gs) will theoretically not even notice). As all of my previous hosting woes have been caused by faulty equipment (and bad service), and since (gs) seems to have solved the problems that come with having your site tied to one specific piece of equipment, in the end I went with MediaTemple (gs). (And although there have been some reported problems with the service, I have not experienced anything but very fast loadtime and well-designed control panels since I have started migrating my sites over).

Every few days I go in and clean out my spam folder. Today there were 399 conversations waiting for me. Usually I have to do the following:

1. Select All 100 conversations on the screen
2. Delete
3. Repeat for every screen

I understand why they make you do this, but it still gets annoying when you have a bunch of screens to take care of.

Today when I went to delete my spam conversations, what do I see?

That’s right. You can now select all of my spam conversations and delete them at once. Or don’t. (I did). The nice thing is that there are new features constantly being added with the intention of making my life easier, something only possible with a web application. Thanks Gmail!

While “rolling some heads” every once in a while is not such a bad idea, I do not think that the Google product line is in as much danger as one might think. The New York Times story referenced earlier (In the Race With Google, It�s Consistency vs. �Wow�) hints at the reason: Although Google might be losing the footrace in terms of numbers when it comes to news, email and IM, they are making up for it in terms of quality and most importantly, the “Wow” factor. They use the type of technology that the techies and Slashdot crowd like (compare Google Finance with Yahoo Finance to see what I mean), and this most-important sector of the market is the ones who are driving tomorrows technology trends. Microsoft, Yahoo and AOL have had years to build up huge subscriber bases with Email and IM, so it is understandable that Google has a long way to go. However, their growth is stagnating and their products are not so attractive compared to what Google (and other similar companies) have to offer. Put it this way: whenever I hear from a friend or acquaintance that they set up a new email account for personal use, 99% of the time it is a gmail account. That is why these numbers don’t really seem to me to be such bad news for Google int he long-term. They do not capture the overall market trend, especially among the market-movers.

So I installed it, as I have done on a number of computers in the past. I then went and tried to run browse the webapp in question (using http://localhost/appName) and received an HTTP 500 error (IE7 politely told me that this indicated a server error, and that the website might be under maintenance).

I tried loading a totally different webapp that had nothing to do with Crystal. Same error. Reboot, same error. Not good.

So I looked in the System Event Viewer, and saw a bunch of the following warnings:

The server failed to load application ‘/LM/W3SVC/1/Root/appName’. The error was ‘General access denied error’

OK. What is that supposed to mean? And why was it killing all of my webapps?

After a little bit of Googling, I found a page by Brett Hill in IIS Insider (March 2005), with one of the subsections titled: Resolving the Error: The server failed to load application…

The most likely cause of this problem is that the DTC coordinator service has not started. If it has started, then look in the Event Viewer for a message regarding a failed logon for the IWAM_servername account.

So I went into Services, and the Distributed Transaction Coordinator service was running. I stopped the service and restarted it, and voila, all of my sites would run again. Hopefully this situation will continue after my next reboot. If it does, the above article has some more advice on ways around this.

I have installed this program a number of times already on other machines without experiencing this error. Anyone know what may have caused it.

Innovative features and things that I like:

• When you approve a story to a certain level (on the sliding scale) the site takes your feedback and automatically adds a new story to match your preferences. Good use of AJAx here.
• Full use of features without creating account – view your history, use your feedback to show you more of what you might like
• All in ASP.net, with tastefully done AJAX (or ATLAS?). Site seems to be operating smoothly, thoughtfully designed.
• I like the feature where you click More and get to view a DHTML onscreen popup showing the text of the entire article. Nice feature.

Unfortunately, that’s it for things that really stand out for me. The Web has a bunch of different sites already that take personal feedback to produce news. Although there is no real stand-out site right now that gives you personalized results, I don’t think that this site is there yet (I know that they are beta, and that they launched yesterday). So a little feedback on things that I didn’t like or things I feel could use improvement:

• Although at first glance, the sliding scale is cool, I find that after using it about ten times it is getting old. When I like an article, I just move the sliding scale up. When I dont like it, I move it down. I don’t really take care to move it to the exact number that might apply (nor am I going to spend the time to figure out if I like the site 2.5 or 3.3). Digg-style voting (thumbs-up or thumbs-down) is much more intuitive and easy to use. Perhaps if you want to expand on this, have four buttons, ranging from -2 to +2. But the sliding scale doesn’t seem so affective right now.
• Since you guys are into AJAX, could I have a drag-and-drop to position the categories the way that I see fit?
• The categories themselves don’t seem to be so clearly defined (or at least the definitions are not intuitive). I am seeing web-design related articles in the Arts category not in Computers and Internet like I would think that it would apply. I know that I could say that I want to see more things in C&I and less in Art. But that is not true, I want to see more web-design and programming…but not more Arts (if that made sense).
• I would like to see more stories on a page. Right now, with the two column layout, I see six stories on a screen when I load up a page. I know that this is similar to Digg (which shows five stories). I much prefer the Newsvine approach (at least 20-30 links). Perhaps in addition to your one-two column viewing options, you could add the option to view just the headlines (allowing more to fit on one page), or the headlines with snippets.
• Although I like personalized news, I am also interested in seeing how other people ranked a story. I know that there are comments, but the more of a sense of community I can see, the more drawn to the site I will be.

Altogether, I think that the general implementation was done well and the site could catch on. However, my initial experience is not enough to make me add this to my list of frequented sites just yet (and with no RSS, it will be easy for it to drop off my radar completely).

1. Absolute: “http://ellisweb.net/” – this will work from any page, and will always point to the proper destination. These can get annoying though. After all, who wants to type out the full html for every page, especially when they are pages on your site
2. Relative: “photos.html” – this will point to the given file, residing in the same directory as your current file. This works as long as you don’t change your directory structure. Ever. Additionally, this can be a bit confusing and easy to mess-up sometimes, especially when you have links like “../../../Articles/123.html”.
3. Relative to the Root: “/photos.html” – this is kind of in between the first two types. It only works on your own site. It identifies the exact location of a file, but identifies the location relative to the root directory of your website. Thus, as long as the photos.html file stays in the same location, this link will always work, regardless of the point of origin. Furthermore, if someone hovers over the link and copies it, it will copy the full path (ie: http://example.com/photos.html).

If you think that the location of a page will change, the third option above seems to have become an acceptable way to link to pages. (And if you know for sure that your site directory structure will never change, then there is nothing wrong with relative links). However, with the proliferation of RSS syndication, using either type of relative link can lead to unforeseeable outcomes.

For example, in Daily Grind 853, Mike has a sentence at the top: “Well, thanks to the site redesign, there’s now a page full of them.” (he is referring to site logos). In this sentence, the words “page full of them” are linked to http://www.larkware.com/linking_images.html. At least, that is the way they are linked if you look at the site directly.

I subscribe to the RSS feed for the Daily Grind, so I initially read this through my Bloglines account. When I went to click on this link, I ended up at http://bloglines.com/linking_images.html, giving me a “404 Not Found” error. Why did this happen?

Examining the source for this page, the link tag is defined as <a href=”../linking_images.html”>. So what happened? Mike used a relative link that works perfectly on his own site. However, when Bloglines (and presumably other online newsreaders as well) served this out, they just published the relative link as it was in the published feed. This caused the incorrect linking situation that I experienced above.

(Sorry to pick on you Mike, I have seen this happen from other sites too. Yours is just the one that prompted me to write this).

The moral of the story: If you publish an RSS (or ATOM) feed, and you want to make sure that your readers will always be able to get through to your links, no matter where they are reading your content, stick with absolute linking.

(And to any programmers working for Bloglines or their competitors – for the sake of your users, could you guys stick some logic into your display that automatically converts relative links in feeds into absolute links?)

Technorati Tags: HTML, linking, Bloglines, RSS, Atom