I don't know how much of this is new relative to other stuff I
have posted about Eazel/Nautilus/etc in the past, but I found it
pretty interesting.
http://www.linuxpower.org/display.php?id=209
> Eazel making its way forward in a sea of change
>
> Interviewed by [21]Christian F.K. Schaller
>
> [22]Eazel is not an old company, yet they already have had a fair
> share of ups and downs. Their entry into the Linux scene was widely
> applauded and their work on Nautilus has been greatly anticipated.
> With that in mind, the downturn (which came as a direct result of the
> slowing US economy) came as a big shock to those inside and outside of
> Eazel. In the middle of this difficult process of change for Eazel,
> [23]Nautilus 1.0 was released. To find out more about all of this, and
> how Eazel plans to put their current problems behind them and move
> forward once more, I talked to [24]Bart Decrem, [25]Darin Adler, Ken
> Kocienda and [26]Maciej Stachowiak.
>
> For those of you who don't know, Darin Adler is the technical lead for
> the Nautilus team. Among other things, Darin's track record includes
> being the technical lead and chief architect for Mac OS System 7. Bart
> Decrem is one of Eazel's co-founders and a member of the board of
> directors on the [27]GNOME Foundation. Ken Kocienda is Eazel's
> Director of Services Engineering.Maciej Stachowiak is a
> hyper-productive code cowboy, Technical Lead of Eazel Services
> Engineering and a member of the board of directors on the GNOME
> Foundation.
>
> Christian: Ok, let's start with the tough stuff. A short while back
> you had to let go a large part of your staff due to delays in getting
> venture capital for your third and last round of funding. How have you
> worked your way through that ordeal?
> Maciej: It's been a tough time around Eazel. However, we managed to
> keep most of the core engineering team intact, and there is a lot of
> energy now. I like the way Bud Tribble, our VP of Engineering, put it:
> "We have a lot of momentum. We've lost some mass. So that means we
> have to increase velocity." And it's been happening - we've been
> achieving things in a matter of weeks that would have taken months
> before.
>
> Bart Decrem: But it's been very painful. We had to let go most of the
> people on the business side of the company, in QE and other areas.
> Even in engineering we lost some really good people. For example, we
> had to lay off several GNOME hackers who were part-time employees.
> When you're laying off full-time staff whose families are depending on
> these jobs, you have to also let go all your contractors and part-time
> people. It was particularly sad that we had to make this announcement
> on the same day that we did the 1.0 launch of Nautilus. Some people
> have speculated that this was some sort of evil conspiracy, but it
> really wasn't: a few days before the 1.0 release, it became clear that
> we would need additional time to secure our next round of financing,
> and we needed to take immediate action to ensure the viability of the
> company.
>
> Christian: As mentioned in the previous question you have yet to
> secure your third round of funding. How serious is this situation?
> Bart Decrem: Actually, we've only had one round of venture capital
> funding. But one thing is for sure: things are a lot different today
> than a year ago, when we closed our last round, or even a few months
> ago, when [28]Ximian secured financing. The reality is, the US economy
> is headed into a recession, dot-coms are going out of business all
> around us, venture capital funds are taking severe beatings and Linux
> stocks are down a lot. So a year ago, investors were willing to bet on
> unproven but potentially huge future markets. Today, investors are a
> lot more risk averse. Even corporate investors, the leading computer
> industry companies that stand to gain a lot from the success of GNOME,
> are much more conservative: they too are laying off employees and
> losing money on some of their investments. Nonetheless, a number of
> investors are quite interested in what we've been building. So we're
> in discussions and pursuing a number of different opportunities, and
> we expect to have news sometime over the next 60 days. But we did have
> to cut our overhead in order to adjust to the new economic environment
> and also to make sure the clock doesn't run out on us while we're
> securing new financing.
>
> Christian: The positive results from [29]Red Hat are probably a good
> thing to bring to the table in your upcoming funding discussion?
> Bart Decrem: Yes! Their financial results are quite impressive: they
> doubled their income year-over-year, to an annual rate of $100
> million, and basically broken even. They're definitely proving that
> there's major business in free software! The reality is that Red Hat
> is the leader in the Linux industry, so what's good for Red Hat is
> good for the rest of us. We're particularly excited to see them moving
> full steam ahead with the Red Hat Network, and roll out their
> subscription plans for that. If we want free software to take over the
> world (and we certainly do) we need to build on the efforts of the
> thousands of hackers that have brought us this far, but we also need
> to have successful businesses that can offer credible alternatives to
> what Microsoft has to offer. And nowhere is that more important than
> in the area of web-based services, particularly in view of Microsoft's
> recent .Net announcements.
>
> Christian: Nautilus 1.0 was recently released and it was met with
> mixed emotions. Eazel has been keeping a low profile in commenting on
> user response, but what are your feelings on the feedback up to now?
> Bart Decrem: Yes, we got our share of flames on [30]Gnotices last
> week. But hey, that's how hackers express their affectionate feelings.
> To be fair, we also got tons of positive feedback from people who
> really love the product. But none of us here at Eazel had the
> emotional energy last week to respond to the flamage. Now, the whole
> point of getting to a 1.0 release is to get the product out there to
> more people so you can get the feedback you need to make the product
> even better. Fundamentally, I agree with much of the criticism we've
> received. But the reality is that Nautilus is a very ambitious
> project, and we depend heavily on rapidly evolving technologies such
> as Bonobo and [31]Mozilla. So we're proud of the product we've
> shipped, and we're listening to the feedback, and we're committed to
> addressing the concerns we've heard over the next few releases.
>
> One of the things we've done to capture the feedback is we've set up a
> mailing list, [32]
[email protected], where people can submit feedback.
> Then we went through all the feedback we got there, and on Gnotices,
> and on our support list, and on IRC, and we digested all that feedback
> and came up with the Top 21 Most Requested Features and Fixes for
> Nautilus. Here they are:
>
> 1. Speed! (includes .jpg thumbnailing performance)
> 2. Ability to more easily move files around (using either
> cut/copy/paste or shelf idea)
> 3. View content of .tar, .rpm packages
> 4. Ability to delete files directly without going through the Trash folder
> 5. Integrated shell features
> 6. Better browser functionality (inc. Download functionality, Mozilla
> related problems etc.)
> 7. A no-frills, lean and mean file manager (no Mozilla, no services)
> 8. Better MIME type support
> 9. Clean quit
> 10. Elegant way to drop app launcher icons on the desktop
> 11. Better keyboard navigation
> 12. Bootstrap installer better recovery after failures
> 13. Sort directories before individual files are listed
> 14. Text on desktop more readable
> 15. Stability
> 16. More previews (HTML, pdf)
> 17. Automount floppies
> 18. SMB support
> 19. Root password popup box
> 20. TLC to Tree View and List View (easy Rename in List View, for
> example)
> 21. Virtual folders not just as bookmarks
>
> (Links to all the feedback reports, and our summary reports, are
> archived at [33]here.)
> So now Darin, Don and the rest of the Nautilus team are doing bug
> triage and figuring out how we can address these concerns. We're
> hoping to come up with a new roadmap over the next week or two.
>
> Maciej: There are many features that people would have liked to see in
> 1.0 which did not make the cut, because we really wanted to ship and
> get something out to users. The many months of all bug fixing and no
> new features were pretty unpleasant. But we're really glad to have 1.0
> out there, because it means we can add a lot of the features that
> users want and that we'd like to see ourselves.
>
> Of course, when I say we, that's not completely accurate, I'm not
> working on Nautilus that much any more. I am focusing more on Eazel's
> service offerings now, particularly Reef, our new services
> architecture.
>
> Christian: Another issue is performance, is this something you will
> continue looking into?
> Bart Decrem: Performance is the number one complaint people have about
> Nautilus, so it's number 1 on our list of priorities. Performance
> actually improved dramatically between PR3 and 1.0, but we need to do
> better still.
>
> Christian: What are Eazel's top priorities for the next releases?
> Darin Adler: Our top priority is fixing any major problems seen in the
> field now that Nautilus has more real users with 1.0. We're revisiting
> some design decisions, and adding the key features that are most often
> requested, like a Windows-like copy and paste for copying files and a
> Scripts menu for running often-used commands on selected files in the
> icon and list view. As Bart pointed out, we're going out of our way to
> solicit suggestions from the GNOME community to drive those choices.
> We're refining details, like making sure that icons don't overlap the
> menu bar if you have one at the top of the screen. We're making
> changes that speed the program up. And we're improving the
> internationalization, making it work better for locales like Japan and
> doing more testing.
>
> Christian: So does this mean that we will not have to wait long for
> the next release?
> Bart Decrem: Give us another week or two to finalize our new roadmap.
> But yeah, I'm dying to see us ship Andy's cool RSS feature, for
> example, which lets you drag and drop dynamic web content, such as /.
> headlines, onto the desktop.
>
> Christian: Eazel has implemented a system in Nautilus with 3 levels
> for different kind of users, will you work with other GNOME hackers to
> implement this system in more GNOME applications?
> Bart Decrem: We would like to see this happen, but it's one of the
> areas where we will really need to engage the GNOME community and make
> sure everybody buys in to the idea.
>
> Christian: One of the cool features of Nautilus is the use of the
> first lines of a text file as its icon, unfortunately this doesn't
> give much useful feedback in the case of XML files like [34]Gnumeric,
> [35]Abiword or [36]OpenOffice. Any plans to have Nautilus ignore/use
> the XML formatting for such files?
> Bart Decrem: Yes. We have plans for pluggable thumbnailing for any
> kind of file. So, for example, you could have an icon for a web page
> and that'll show up as an actual live thumbnail of that size. And
> since icons are resizable, you can make that as big as you want, and
> drop it right on your desktop. Or if you're looking at source code, we
> could offer syntax highlighting. So we have a bunch of interesting
> ideas in this regard, and some of those we'll be doing ourselves, but
> we've tried to architect Nautilus so that it's easy for people to
> contribute these sorts of things.
>
> Christian: What do you think are the main innovations you have put
> into Nautilus?
> Bart Decrem: Well, from a development perspective, Nautilus is the
> first GNOME app to reach the 1.0 stage that's implementing hot new
> GNOME technologies such as GNOME-VFS and Bonobo. One of the things
> that's awesome about GNOME-VFS, and another example of how we're
> bringing the Internet to the user's desktop, is that with GNOME-VFS,
> you can browse local files or networked files. So with Nautilus you
> can access FTP files, WebDAV directories etc. And as GNOME-VFS gets
> adopted by other GNOME apps, people will be able to store their files
> directly to the web. One of the ways we're using Bonobo is to embed
> Mozilla. As you may know, the trick with building a web browser these
> days is not so much to build the basic browser, but to deal with all
> the weird ways the major web sites screw up the code on their sites.
> So it's really hard to build a browser that will actually properly
> display all the popular web sites. So rather than try to reinvent the
> web browser, Bonobo lets us just plug in Mozilla.
>
> From a user perspective, there's also a ton of innovation in Nautilus,
> including sizable icons, the user level concept and the use of
> emblems. The emblems are a great way to flag your documents and help
> you find the important stuff, but you can do a ton of really neat
> things with them. For example, one of the cool features in Nautilus is
> our full-text search technology, Medusa, which indexes your entire
> hard drive (it's kind of a combination of Find and Locate). So you can
> do a search and say "show me all my files that have the Documents
> emblem" and then you can bookmark the result of that search. What
> you're getting is actually a dynamic page. So next time when you pull
> up that bookmark, you'll see all your documents. It's kind of like the
> virtual folders concept in mail clients like Evolution. A future use
> of emblems might be that you can flag a folder as Important, and then
> that folder will automatically get backed up on the Internet at
> midnight each day.
>
> Christian: Eazel recently hired [37]John Harper and [38]Georg Lebl
> (the main authors of [39]Sawfish and the Gnome panel respectively),
> was this just a move to support an important component of the GNOME
> desktop, or do you have some plans for extended Nautilus integration
> which you want to see happen?
> Bart Decrem: We hired John because we were really impressed with his
> work on Sawfish but also because we are very interested in the apps
> that affect the core desktop, and obviously the window manager is an
> important part of that. Having John on our team means that we don't
> have to worry about the file manager and the window manager fighting
> over who gets to draw the desktop, for example. More importantly, John
> is a really great hacker, and he's kind of a one-man fire squadron
> who's always available to help us tackle particularly tough technical
> challenges. Similarly, with George's help, we were planning to do some
> really cool integration between the file manager and the panel.
> Unfortunately, George was a part-time employee here, and we had to let
> him go last week, but we hope that he'll stay involved if we give him
> enough Free Beer.
>
> Christian: Tell us about Eazel's services, what do they currently
> include and what kind of things can we expect in the future?
> Maciej: Right now, we have two primary services, Online Storage
> (sometimes referred to as the Vault), and the Software Catalog. Online
> Storage provides you with networked storage that looks just like a
> normal directory in Nautilus, but can be accessed from anywhere in the
> world. Because we built on WebDAV and other open standards, you can
> get at your stuff even from Windows boxes, or from any normal web
> browser.
>
> Developing these foundational services was Phase 1 of our services
> development plan. We learned a lot of lessons from the way we
> implemented services UI, which was a combination of HTML viewed in
> Mozilla and native widget code implemented in C. Phase 2 will include
> a new way of deploying services, and a number of new services based on
> the foundation ones we have now.
>
> Christian: How do you plan on handling development of the services
> from here on? I think something called Reef was mentioned earlier in
> the discussion.
> Ken Kocienda: In our first attempts to develop services, two tracks
> evolved: developing components written in C and distributed with
> Nautilus, and web pages delivered to Nautilus and rendered by the
> embedded Mozilla component. The former provided us with the direct
> access to core Nautilus functionality that we needed for things like
> showing remote, online storage files with the standard icon and list
> views, while the latter gave us the flexibility to evolve UI and
> interaction designs quickly without affecting Nautilus development.
> Now, looking beyond Nautulis 1.0, we want to merge these two tracks.
> This is what Reef is about: rich functionality delivered in a
> flexible, dynamically updatable format. Reef components are
> self-contained programs that Nautilus downloads off the network and
> knows how to run in its local environment.
>
> In some ways, they can be thought of as applets, but unlike applets,
> they are developed to integrate very closely into the environment
> provided by the host container, in this case Nautilus.
>
> Maciej: So Reef is the project to implement our new services
> architecture. It is based around the concept of Service View Bundles
> (SVBs), downloadable self-contained bundles of script code, images,
> and other resources (for example html or glade files). These bundles
> are going to be signed by Eazel and will run inside Nautilus.
>
> They'll talk back to a server using open protocols like XML-RPC and
> SOAP (ultimately our hope is to migrate to the ultimate XML Protocol
> standard established by the W3C, in the meantime we're staying
> flexible).
> The great thing about this approach is that it will make services much
> easier to develop, as well as snappier, more fun to use, and better
> looking.
>
> Christian:The new services are being developed using [40]Python, any
> specific reasons for this choice?
> Ken Kocienda: Python is a good initial choice for several reasons. It
> is a quick prototyping language that scales up well as a project
> becomes larger, it has a wealth of libraries, and it has very good
> bindings to [41]GTK+. Also, Python's rexec and Bastion features give
> us the tools to address some of the security concerns raised by
> shipping code around the Net that executes on local systems. However,
> with Reef, we're not focused solely on Python. We want to make it
> possible to develop in other languages as well.
>
> Maciej: Even though we are starting out with Python, we are designing
> things to make it easy to add multi-language support later; we're
> especially interested in supporting [42]Perl, [43]Guile, Rep and
> Javascript.
>
> We really wanted a language that was easy to write in, popular,
> maintainable and portable. We also wanted good support for things like
> networking, GTK+ and GNOME bindings, and integration with C, as well
> as an active user and developer community. Python fits the bill
> perfectly. It has some of the nicest GTK+ bindings out there, and it's
> really easy to write in for both beginners and experts.
>
> Python/GTK+ code is really the middle ground we wanted between HTML
> and native widget code written in C. It's lightweight, portable, and
> easy to deploy, but it can give a rich interface, better UI
> performance, and a native look and feel.
> We also wanted something that would be easy for other developers to
> use to build their own services.
>
> Christian: So in other words if people want to see a certain service
> provided by Eazel they can implement it themselves? :-)
> Ken Kocienda: Yes. But, we also see this as a way for people to make
> their own Nautilus views, without ever intending to ship those over
> the network to anyone else. Part of the goal is to lower the bar to
> extending what Nautilus can do.
>
> Maciej: People who want to provide a service themselves (whether for
> money or for free) can implement it themselves. We're trying to do
> this as a true community project, directly in public CVS from day one.
> We just recently made an announcement about the project and set up a
> public mailing list and IRC channel.
>
> We've also thought about a lot of directions to take the Reef
> architecture, including a lightweight Reef viewer that would work on
> small devices like the iPaq, and one that would plug into the
> [44]Evolution Executive Summary.
>
> We want it to meet everyone's needs for a lightweight, flexible
> application deployment framework.
>
> Christian: Who will represent Eazel on the GNOME Advisory board?
> Bart Decrem: Bud Tribble, he's our VP of Software Engineering. Bud has
> a very interesting background. He was the first manager of the
> Macintosh team at [45]Apple, then he went away to complete his MD/PhD
> in neurophysiology (man, I can't even spell that:). He went on to be
> the cofounder, along with Steve Jobs, of [46]NeXT, where he was
> responsible for NeXTStep. Then he went to work at Sun where, most
> recently, he was the CTO of [47]iPlanet, the Sun-Netscape alliance.
>
> I think the GNOME Foundation's Advisory Board has some really awesome
> people on it. For example Rob Gingell from Sun, he created the dynamic
> linking mechanisms now prevalent in UNIX implementations, worked on
> the ELF object file format, and led OS development in the areas of
> memory management and multithreading. He's the CTO for [48]Sun's
> software group.
> So I really look forward to finding ways to tap into some of these
> guys and get their help in really mapping the road ahead for GNOME.
>
> Christian: What kind of things do Eazel hope to see changed/get added
> in GNOME for the 2.0 release?
> Maciej: One thing I'd like to see is to push some of the technologies
> we've developed for Nautilus into the project as a whole - things like
> user levels, some of the custom widgets we've developed, our themable
> icon framework, and so on.
> Another thing I'd like to see is to maintain the discipline the GNOME
> project has had for GNOME 1.4, and set and stick to a schedule.
>
> Published March 30, 2001
>
> References
:
> 21. mailto:
[email protected]
> 22.
http://www.eazel.com/
> 23.
http://nautilus.eazel.com/
> 24.
http://www.advogato.org/person/bartd/
> 25.
http://www.advogato.org/person/Darin/
> 26.
http://www.advogato.org/person/mjs/
> 27.
http://foundation.gnome.org/
> 28.
http://www.ximian.com/
> 29.
http://www.redhat.com/
> 30.
http://news.gnome.org/gnome-news/
> 31.
http://www.mozilla.org/
> 32. mailto:
[email protected]
> 33.
http://lists.eazel.com/pipermail/nautilus-list/2001-March/002280.html
> 34.
http://www.gnome.org/projects/gnumeric/
> 35.
http://www.abisource.com/
> 36.
http://www.openoffice.org/
> 37.
http://www.advogato.org/person/jsh/
> 38.
http://www.advogato.org/person/vicious/
> 39.
http://sawmill.sourceforge.net/
> 40.
http://www.python.org/
> 41.
http://www.gtk.org/
> 42.
http://www.perl.org/
> 43.
http://www.gnu.org/software/guile/
> 44.
http://www.ximian.com/apps/evolution.php3
> 45.
http://www.apple.com/
> 46.
http://www.apple.com/enterprise/
> 47.
http://www.iplanet.com/
> 48.
http://www.sun.com/
--
Gerald Oskoboiny <
[email protected]>
http://impressive.net/people/gerald/