Message: previous - next
Month: June 2012

Re: [trinity-devel] Note on udev merge into systemd

From: Darrell Anderson <humanreadable@...>
Date: Thu, 28 Jun 2012 14:09:06 -0700 (PDT)
>   You look across the board at the "progress" and I
> began to wonder just how much has been 'lost' to progress. kde3, Qt3, gnome 2, hal,
> udev, Quanta+, now gimp 2.8[1] and the list goes on and on. Good working
> applications that met their need well, abandoned well before any well thought out
> replacement was in place and ready for end-use. I understand the need to
> innovate and the need to change, but change for "change's" sake has never been a good
> thing.
>   One of the pivotal, and correct, arguments against
> the adoption of open-source for a personal or business desktop -- is you cannot rely on
> what works being here next month, much less next year. The cost for business
> and governments to "re-train" for the next greatest desktop or to "re-tool" and
> implement the next greatest "backend" (to whatever) simply prevents adoption of
> Linux in most cases.
>   Can you imagine the chaos to business if they had
> adopted kde 3.5 only to have 4.0 forced on them in May/June 2008? It's 2012 and that
> desktop still struggles for usability. Gnome3 almost as bad a transition.
>   You think about the work we do to keep up with all
> the next gee-whiz ideas of the dependency packages and imagine business trying to
> justify keeping 50 million people/boxes on the same page.
>   I think if open-source has learned nothing else, it
> has learned how NOT to manage a desktop transition in the kde4/qt4 experience.
> Though gnome3 does bring that into question.
>   That's why projects like TDE are so important. The
> provide the long term stability for a very usable desktop as a choice that makes
> sense to everyone, including business and government.
>   Yes, windowmaker and fluxbox have been around
> forever, but just try to teach a secretary to use them and try to justify managing and
> installing all of the helper apps required just to provide the basic functionality
> that something like TDE provides. It just can't be done economically.
>   In my mind this is the very reason we have for
> putting the effort into TDE and doing it right. Open-source can make a fantastic desktop and
> provide a valid alternative to the proprietary offerings, but only if people
> can rely on it being here tomorrow and working the way it did last month,
> last year, etc.. I am an open-source advocate, but I do take note of the value
> that stability provides every time I see xp boot. All lessons we do well to learn.
> footnotes:
> [1] this is more about the idiotic open/save/import/export
> dialog changes that forces and intermediate save in .xcf. Another great example
> of 10 years of progress developing a flexible open/save dialog tossed out
> the window on the seeming whim of a couple of people.

I share the gist of your sentiments. Yet I realize this is a complicated subject.

I want Trinity to remain static in certain ways yet also change. I want the basic desktop model we use to remain intact. I don't want to run a caching database backend to check my email or run kalarm. I want others to have the desktop they want but I don't want social networking or semantic desktops rammed down my proverbial throat. I'm not interested in a tablet or cell phone metaphor on my desktop computer.

Yet as a team we have to ensure the Trinity desktop model remains functional, such as the hardware abstraction support hacked by Tim to replace HAL.

I grow weary of the geek mentality of change for the sake of change. Redecorating the house, rearranging the furniture, shampooing the carpets, or painting a few walls is one thing. Placing the house on stilts and replacing the foundation every couple of years is a tiresome process and silly.

I grow weary that many distro maintainers blindly adopt whatever technology some geek at Redhat announces.

I am aware that the human species never moved forward by those who only embrace the status quo. We become better as a species by those who think outside the box. New ideas challenge us to think differently. My concern is not new ideas but new ideas being adopted blindly and with a fan-boy attitude rather than actually proving an idea. Worse, is when new ideas are adopted and those who prefer a different way are ignored. My peeve with the latest desktop models is not that I refuse to change, but that I evaluated those models, decided I prefer the traditional model, and am told to change or move out of the way.

As somebody who has provided technical writing services most of my adult life, I have worked with good and bad engineers. I'm referring to their engineering skills and not their personalities. Some focus on function before form and others the opposite. The former group can be frustrating because they tend to focus only on function. Once they provide the function they don't care about everybody else with the form. The latter group is frustrating because they create a pretty product that is functionally useless or buggy. Rare is the engineer/programmer who sees both the proverbial forest and trees.

One challenge in free/libre software is most of the developers are geeks first and not users. Geeks tinker. Most people do not tinker. Geeks are DIY (Do It Yourself) people. Most people are not. Geeks design software for geeks. This is somewhat like the problem of certain cars where changing a spark plug requires lifting the entire engine. The geeks like hoisting the entire engine to change a spark plug. They like convoluted command line options and parameters. They like opening text editors to modify a configuration file rather than provide a GUI control. Most people don't. Most free/libre software is designed by the geeks for the geeks. I don't have an answer to my observations. With respect to the health of Trinity, some of us in this project need to be vocal as users and not just as geeks.

In part this type of discssuion reduces to the fact that coordinating human action is challenging. People want different things and that applies to computers and desktop models. I don't like the direction the newer models are being taken by developers, but I don't want to stand in their way. Conversely, I don't want them to stand in my way to use a traditional desktop model or to be told I am wasting my efforts sustaining the older model.

I want very much to see Trinity succeed for several years. I don't like what is happening to the computer desktop model and therefore am not warm and fuzzy about my options in that area. We need to keep the underlying hooks functional to keep Trinity useful, but I want and hope we leave the overall model as is.

Okay, you can have the soap box back now.