How difficult is it getting a Hardware Test Kit going ?

Firstly, the above thing does not really exist. At least, not to my knowledge. There are variants of the concept which are built around “certification” or “compatibility” testing – more in terms of providing a metrics driven output of interaction of the Operating System with the hardware. But that is not what I meant. One of the reasons for this post could be a small mail-thread. Another could be that given the fact that “non enterprise” Linux operating systems by the very virtue of their fast release cycles don’t have the luxury of having Hardware Certification sites, it would not be too insane to provide the tools (or a testing kit) to the users to figure out where they stand. A last reason could be this too.

In very simple terms the idea is to provide the user with a tool (and perhaps a few plugins/addons) that would allow the user to download the following details :

  1. the details about the hardware they propose to test the OS against (eg a notebook/desktop one is going to buy)
  2. the details about the extent of hardware support provided by the OS of choice

So, if for an use case I want to buy a Dell Latitude 630 and want to run Fedora 7 on it, using the tool I can take a shade-better-than-guesstimate as to how the hardware and the OS intersect. Having said that, what are some of the blockers in this ?

  • getting the hardware information to be released in a format by the vendor that can be used by the tool across any Linux OS
  • getting the Linux OS to create an “extent of hardware support” for that particular version

The second  item might not be all too difficult  given that an OS release does factor in to a large extent the  hardware that  it will end up  being installed on. Being based on a sizeable section of upstream code also allows that information to be captured. The trick is to collate the information and make it available in a form that allows a comparison however rudimentary. Which of course brings us to the first item – that of getting hardware vendors to  release the information in a form that can be useful. All models have technical specifications, it might a bit of extra effort to ensure that they are properly structured to allow comparison. The extra effort tradeoff would surely be an increased uptake in the desktop/notebook space by folks who are keen to have Linux on their old and new desktops/notebooks.

A typical UseCase would be something like – I want to buy a Wipro WIVNB7711 notebook and plan to run Fedora 7 on it. I have the tool downloaded on my existing/current box, added the database for Fedora 7 and then downloaded from Wipro’s specification site the specifications for the model. I run the specifications against the OS database and visually see what would be the result even before I buy-install-do_the_round_of_LUGs. Extend the same to printers, scanners and other peripherals and would get to see the impact of working on such a tool.

The brand that is Fedora

There is a nice mail (currently seems to be more of a tentative push) from Mike McGrath about “the brand that is Fedora“. The emphasis is mine.

For a long time now, Fedora folks have not really been talking about the changes in the infrastructure, look-n-feel, the user space focus, the investments into desktop and laptop improvements that has been taking place save the infrequent LWN article. Given that Moonshine has been a sort of the other side of the flex point release this is as good a time as any to get some brand love in place. This would at the end of the day also enable all to arrive at a set of guidelines into:

  • Fedora values
  • How to use the name Fedora
  • The design elements that make up the Fedora Project and are reflected in the artwork, wiki, doc stylesheets etc
  • User Interface guidelines
  • Grammar guidelines
  • Abbreviations
  • Trademarks
  • Typefaces
  • Colors (color palette)

A brand style guide does go a long way in making a project messaging and context “sticky”

Time to form a body for Open Source within India …

This cutesy mail about “organizational reflex” set me thinking whether one should strive towards creation of a national level industry driven body for Free and Open Source within India. Right now, the ones who shoe horn this function include MAIT, NASSCOM and a few scattered bodies including NRCFOSS. The problem in the current scenario is more of communication rather than coordination. What would be well served in terms of functional goals if such a body was formed might include (but not limit itself to):

  • being the primary point of contact for consultation related to standards, industry information. This would enable Government, companies and the media to be what in cliche-speak is called be on the same page
  • be the national coordinator of events around FOSS especially those related to business advocacy
  • align with other FOSS organisations locally and/or globally to increase the quantum of FOSS contributions from India
  • to help in preparation of reports, guidelines and processes that would facilitate migration to FOSS
  • to help improve local businesses around FOSS

Such a thing has been discussed threadbare in the past and most of the times I guess it has ended in failure with splinter groups working towards a select menu of the above possibilities. I tend to think that given the quantum of commercial uptake of FOSS products and services in the corporate and especially in the government segments, this would be as good a time as any to take some steps in this direction.

I pushed a few tentative mails to a select few folks about this – would blog about the responses or the thought processes as and when they decide to respond :)

Suggestions about getting attention from Airtel Broadband in Pune, Maharashtra

The line at the house has been down since 12th of Sept 2007. I have raised tickets with Airtel with the following scenario:

1. I raise ticket
2. Airtel Customer Support promises visit of technician by 1300
3. At 1330 I call, get visit promised by 1430
4. At 1430 I call, get visit promised by 1700
5. At 1715 I call, get visit promised before 1800
6. At 1815 I call, Customer Support tells me that technician had called me (?) and provided feedback that line was ok, so ticket gets closed
7. A new ticket gets created and we are back at Step 1

The above steps have been happening since 12th and I am really at a loss (not to mention at the end of my patience) as to how to get this fixed. Short of chucking Airtel Broadband out that is.

UPDATE: Since 18th Sept 2007, thanks to a nasty mail to the nodal office, the line (and the internet link) has been up and running

The desktop question …

Aaron writes about “Desktop Linux” and calls it a rant :). It was worth the trouble I took to read it on the phone browser – the internet connection at home is down and Airtel won’t surely get a cookie from me.

The game of Linux on the OEM desktop is primarily driven in two ways: Preloaded or Drop-in-the-box. In most cases, the OEMs who decided not to be shipping preloaded Windows on a few select OEM models used FreeDOS instead. Over the years, some of them also provide a very functionally crippled version of Linux. However, there is a large percentage of them like Dell, HP, HCL, Wipro who either preload Linux or provide drop-in-the-box. The fun bit about doing desktop linux is that the hardware SKU is a small bit faster moving than the server platform. This page would provide an at a glance view of how the OEM market is going to be pertaining to Intel. An inevitable fallout of a fast moving hardware SKU is that slapping on a base Operating System that is a bit dated would be counterproductive. Add to that the inevitable cases of codecs, binary firmwire and it is a fine soup.

A way out for Linux distributions aiming to be preloaded on to desktop hardware is to diligently work with OEMs and ODMs at all levels to check, re-check and check again the functionality of the base OS with respect to the target hardware. While this was a well honed practise in server space, the desktop niche did start to receive base OS engineering love only with the predominant uptake of Linux on the desktop for more than SOHO usage. So, it is not really unusual to see more love being showered on the desktop space in terms of visions and achievable roadmaps like these. But in an ideal world it would take much more than distributions to be interested into desktop enhancement bits. It would really require an extensive formation of a community (oh such an abused word).
The hardest bit is to convince the IHV/OEM to invest heavily into engineering and code-push that makes the hardware of today and tomorrow not only JustWork (TM), but provide a sustaining engineering investment. The Red Hat Global Desktop is just one of the steps along these lines. The thing about a Linux desktop is that it is significantly user experience driven than driven by performance specifications. So, while performance focus would lead to specific FOSS projects polishing up the code base, an user experience focus would require an all out effort based on a good plan. The perception shift is driven by code contributions at all levels of the OS : the kernel, the desktop environment, the IM and IRC clients, the browser and mail clients, the artwork and the bling bits which get talked about. One of the best things that could have happened to a Linux desktop is that it stopped being compared to Windows – one hardly gets to read about “this desktop is so similar to my Windows desktop” and that’s a good thing when one shifts the focus from being relevant in the perspective of competition to being relevant in itself. In effect, the question could well be – what is the metaphor that we now for lack of a better word call a Desktop ?

Control, feeds, bikeshed and a heavy mail archive

Exhibit A, B and C are the ones which are of interest (and somewhat awe). Since I started reading it from somewhere in the middle and kept switching between the two lists, I am not the right person to sum it up. However, on recurring point in one of the threads is the issue of “control” on Planet GNOME. With the disclaimer that this blog does not get syndicated on that planet (and I don’t think it meets the requirements to get syndicated) I do have a few things to say.

When we started off this planet, the quirkiness quotient was derived from Planet GNOME. In fact, I recall that the now-not-so-young-boy who popped the question pointed me to p.g.o saying “can’t we do something like that ?” We did and even though on and off we have heard the remark that “with only that tiny rag tag bunch of feeds you call it Planet FLOSS India” – we have sustained. The single thing that keeps it going is the trust that there’s no heavy handed censoring/editing/policy driven decisions for PFI. There have been decisions for including or removing feeds – but that has been driven by the degenerating nature of feed content even with the slack that is normally given.
The exhibit threads seem to suggest that such a thing exists for p.g.o. Not being an “insider” into the world of p.g.o, I don’t really know. What I do however know p.g.o has for a long period of time been my window into a desktop environment and developer cabal (I mean that in a nice sense). And I enjoy the occasional ramblings into spaces that are not really relevant to the whole “GNOME” thing.

What bothers me is not that there is a huge thread going along these lines – but the fact that a piece of web real estate which is considered the fun side of GNOME is now being dragged into bikeshedding.