If you’ve upgraded to Firefox 2.0.0.12 you probably noticed some changes in the results brought by searches from the top-right corner Google search plugin.

Basically:

– The result page shows a Linux Mint logo instead of a Google one.
– The layout of the page is a bit different (ads are on top instead of being on the right).
– There is no access to advanced Google features (account, cached links, similar pages..etc).

The reason it is different is because instead of using the default plugin we now distribute our own and take advantage of a Google Custom Search Engine. The reason it it is different from the default Google search is because Google doesn’t offer the same features to Custom search engines as it does when searching directly from google.com. The reason we changed from default google to a custom engine is because it generates a lot of revenue and this single plugin could potentially make Linux Mint into a company which actually hires full-time employees.

The highest single source of revenue for Linux Mint isn’t the donations, it isn’t ads on the website, it is the default start page in Firefox. This simple search plugin is estimated to generated from 2 to 40 times more money than the start page itself. If this proves to be the case we’ll be in a position to grow and to achieve things we never dreamed off before. At the moment Linux Mint is maintained by me on a 2.5 hours / day basis with the help of a team of volunteers who are paid nothing. The income is similar to a single small salary. With this plugin, and although you may not enjoy the same level of comfort as when searching with the default Google layout, you contribute to making Linux Mint grow and we should soon see the day where our distribution gets maintained by a team of people working full-time on it.

For most distributions the business model is to sell associated products (Powerpack, Enterprise Desktops, Server offers) or services (Professional Support..etc). Although this allows the distributions to hire full staff employees it also takes a lot of resources and most employees end up working on these associated services/products instead of focusing on the distribution itself. If we find a way to finance ourselves by clever little changes like this one we can grow into a medium sized company with full time employees working ONLY on the distribution itself. As you can see this is a very interesting perspective.

For people used to using Google Images we included a second search engine so that they could achieve the same things as before. For people used to cached links and used to access their google account there is a workaround and it is possible to revert to the default Firefox plugin. Of course this is a pity and we don’t encourage you to do that (since this is the best way you can contribute to Mint and it costs you nothing) but it’s your desktop, not ours, and we always believed you should have a choice. So if you really want to revert back to the default plugin, follow the steps below:

Go to https://addons.mozilla.org/en-US/firefox/addon/3682
Install the addon and restart Firefox
Go to google http://www.google.com
Right click in google search bar and select “add to search bar”
Left-click arrow at left side of firefox search combo
Select “manage search engine”
You should find google at the bottom of the list
Move it to the top

By doing so your search results will use the original Google layout again (cached links, etc..) but hey everytime you make a search you could have made us grow at the same time.

It’s good to have a choice and it’s good to know why you’re going one direction or another. I’m sorry we rolled out this as an update without informing people. Hopefully this blog post will make things clearer and if you’ve been surprised by this new plugin you’ll now be able to understand how and why you want to keep it or change it back to the original plugin.

I’ll talk about Mint, as a Republic of Ireland registered company, in a separate blog post.

Clem.

After some investigation it looks like a simple call to apt.Cache() from the python apt module is responsible for the high memory usage of mintUpdate. Without this call mintUpdate uses 4.6MB of RAM. As soon as the module is called the memory usage grows to 47MB. The python interpreter doesn’t free up memory and an additional 47MB of RAM is reserved every time mintUpdate refreshes (every 5 minutes by default).

Of course mintUpdate relies on the apt module to get the list of available packages so we’ll need to find some decent solution to replace this call. Either we’ll manage to force python in freeing the memory used by the module or we’ll rely on APT directly without using the module.

Anyway, the next release of mintUpdate should use around 5MB of RAM and this shouldn’t grow over time.

Note: This comes as excellent news as I was personally worried that we would have to rewrite mintUpdate from scratch using another language. The java-gnome bindings lacked popularity and were not even packaged for Debian. Perl kind of scared me a bit when it came to OO-threaded development and I wasn’t particularly amused at the thought of using C++ (although I must say I was quite impressed with how Eclipse handles C++ and with the quality of the Gtkmm API). Well anyway, more on this later.. and hopefully we’ll stay with Python, at least for already existing Mint apps.

Edit: Apparently python won’t let us force any garbage collection. However the apt module can be replaced entirely by calls to “sudo apt-get -s upgrade | grep ^Inst” and “sudo apt-get -s dist-upgrade | grep ^Inst”. We’ll even save calculating dependencies ourselves…

Clem

If you’ve been reading Husse’s newsletter you probably know it by now: Don (aka Exploder) is joining the team and replacing me as the person responsible for testing our ISO images before they get released.

The first advantage of this is that we’re going to achieve better quality and have less bugs slip through. Exploder is much better than I am at finding bugs and when he’s happy to see something released as I do the releases anyway I’ll run my own tests as well on top of it. So the BETA or STABLE ISO which ends on our mirrors will have been tested by the maintainer, by Exploder and by me 🙂 Of course there’ll always be bugs, don’t worry about that 🙂 But at least most of them will be caught prior to the release.

The second advantage of this is that it frees me up from something which required a lot of time. So I can now focus even more on the project, the main edition and the tools we develop.

As Ubuntu is getting ready for an LTS (Long Term Support) release in April, our next release will focus on long term stability rather than on the innovation of a lot of new tools. MintBackup is coming in but don’t expect the same level of innovation as in previous releases: Linux Mint 5 is going to be a boring release 🙂

What we want to achieve before this release is a better vision of how we work, how we define our editions, our tools, how we look at localization and in a very general way how we can improve not the Linux Mint desktop but the project itself and its organization. For instance, one aspect of the upcoming LTS Ubuntu base is that Mint 5 will get 3 years of security updates. This is an opportunity to create an enterprise desktop and a solution which can please companies. So we want to look into that of course. What would be the difference with the main edition? How would we go and conciliate development and updates between the new releases and the LTS one?

If we could manage to have this LTS release declined in two editions, one aimed at companies interested in stability, security, networking and another one similar to what we have now, aimed at individuals who want an elegant desktop. If we could manage to keep our 6 months release cycle, but at the same time still maintain these two LTS releases for another 3 years, and not only by providing Ubuntu updates but by also providing updates for important user tools (OpenOffice, Firefox, Thunderbird, Evolution, etc..). If we could manage to do that, we would be able to keep going on the way we are now but we would have a very interesting option for both companies and individuals. The goal here of course would be to accelerate the migration of companies to Linux and to place Linux Mint on the enterprise market. A 3 years release cycle for LTS is also very reassuring for individuals, especially if they come from Microsoft Windows and if we manage to upgrade user-level applications.

We’re also dropping minor revision numbers for our releases, so the next release won’t be Linux Mint 5.0, it will be Linux Mint 5. The codename hasn’t been decided yet.

If we look at our past releases we can see how the focus went from one area to another. After Bea was released we had a great Ubuntu base with our own tweaks and optimizations and with the necessary codecs and package selection. The focus changed from the system configuration to the development of tools in Bianca and that’s basically what we’ve been doing until the release of Daryna. Now that an LTS release is coming up and also because we’re happy with the quality of our current desktop we want to shift the focus and strengthen the project itself, make our tools more robust, revise our software selection, ensure a better level of localization, produce more documentation, work on our image and our offerings (not in terms of product as we’re not “selling” anything but in terms of what we can offer and to which audience), increase our efficiency in how we do things and let our community grow with us staying close to it.

So you see the list of planned changes for Linux Mint 5 has never been so small and at the same time we’re preparing for one of the most important releases we ever had. This time it won’t be only about the quality of our desktop but about how the project itself managed to grow with its community.

This is a very vague vision of course and as always your feedback is welcome. What we know so far about Linux Mint 5 is this:

  • It will be based on an LTS release (3 year security updates from Ubuntu)
  • We’re hoping to give it 3 years user-level application updates (OpenOffice, Thunderbird, Firefox, all important user-level apps), to maintain it, and while we’re developing other 6 months span releases to give it as much attention as the latest release gets.
  • Most tools will be translated in various languages and get minor improvements.
  • Amarok will be replaced with Rhythmbox (this might change).
  • mintBackup will be introduced.
  • We currently have a Main and a Light edition. It’s not decided yet how or if this will change but we’d like to introduce an enterprise desktop for this LTS, this might merge with Light into a DVD special edition with a radically different software selection.
  • A user-guide will come with the release and we’ll hopefully have that guide translated into many languages.

Right, enough said.. tell us what you think and how you see things happening.

Clem

PS: I didn’t realize I wrote so much.. sorry for the long blog post 🙂