tecosystems

Why It’s Important to Fix the Solaris Shell

Share via Twitter Share via Facebook Share via Linkedin Share via Reddit

I’ve complained ad nauseam about the default shell within Solaris, which is both antiquated and primitive compared to newer shells, with little impact thus far. Most of the folks from Sun I speak with agree that it’s unfortunate, and that it should be fixed, but after several years of complaints nothing’s changed. The usual response is either a.) we can’t because of obscure booting or compatibility reasons, or b.) it’s easy to change it to bash, zsh or something less than several decades old. Either way, I think it’s hurting their ability to recruit new users to Solaris – particularly those coming from Linux.

For example, here’s something from today’s Statcounter weblog:

adsl-[IP REDACTED].dsl.sntc01.pacbell.net ([NAME REDACTED])

California, San Francisco, United States, 0 returning visits
Date Time WebPage
February 20th 2007 12:25:40 www.google.com/search?q=solaris bash word erase doesn%27t work&ie=UTF-8
tecosystems » Solaris-ing the New RedMonk Gear

It’s possible, I suppose, to interpret this differently, but in all likelihood this is someone Googling to find out why backspace doesn’t work on Solaris as it does on Linux, OS X, or Windows. The post they ended up at (I’m number five in that particular query) describes my distaste for the shell, but only indirectly how to set a user up with bash as an alternative. Based on the query, it seems safe to assume that this particular user appreciates the no-backspace “feature” about as much as I do.

What does one user matter? It’s easy, after all, to ignore me because I’ve publicly stated that I have no intention of migrating RedMonk’s business to Solaris, as I’m comfortable on Linux (Ubuntu, more specifically). Even if I hadn’t seen similar queries in the past, however, and I have, I’d care if I was Sun because the name I redacted was – unless it’s someone with the same uncommon name – one of the co-founders of a startup that just about everyone’s heard of.

I know it probably seems spectacularly unimportant to the Solaris engineers, but even little barriers to entry like this can deter would-be Solaris converts. Documentation, installation, and small decisions like the default shell can seem trivial when weighed against some of the technical differentiators within Solaris, but the historical popularity of Windows should be evidence enough that the volume market is more interested in ease-of-use than superior technology. Not learning from that would be a shame.

Disclaimer: Sun is a RedMonk client, Canonical – the corporate entity that funds Ubuntu – is not, although I will be speaking at UbuntuLive Day during OSCON.

13 comments

  1. Actually Stephen, many of us OpenSolaris engineers see these little barriers to entry as being extremely important. And we’re in the process of addressing those including the default shell for root or for new users. It’s a priority now which is not something I could say a few years ago or even a year ago and I think you will see positive change in this space in the coming calendar year.

  2. […] Sr. O’Grady (who it was a pleasure to finally meet last week) has a long post on why the Solaris default shell is silly. It is a good post, and worth reading if for nothing other than his anecdote about who is having problems with the shell. I would summarize it in two sentences: Mom told me that first impressions are important. Did the Solaris people not listen to their moms? (This of course holds true for legal writing too…) […]

  3. I half expected this to be a mirror of a similar problem that’s occuring in ubuntu at the moment.

    https://launchpad.net/bugs/61463

  4. David: there are undoubtedly some – like yourself – that understand both the issue and its importance. i had the privilege of meeting with a bunch of the more progressive minded Solaris devs while out in CA for the analyst conf. that said, there is certainly a traditionalist type of developer that sees little if any need for the change, and is overly (IMO) concerned with the risks.

    anyway, i hope you’re right: i’d love to see some movement on this soon.

    Stephen: yeah, i’ve been bitten by that myself. my usual fix is just call bash to run the script, i.e. bash ./runlime.sh. that’s worked for me, but less than ideal, i agree. particularly b/c most folks won’t know to do that.

  5. The issue with the Solaris shells is currently being worked on – see http://www.opensolaris.org/os/project/ksh93-integration/

  6. When my team’s unix account was set up on solaris, the first thing I asked the admins to do was set up bash as the default shell. They asked me if I was sure everybody on the team was ok with that and I replied “Absolutely”. That was 2 years ago. Nobody has mentioned ksh in all that time.

  7. David Comay, I’m confused by your wording. You’re among fellow engineers; please speak engineering, not marketing. Does “a positive change in this space in the coming calendar year” mean that backspace will work out of the box by 2008?

  8. Scott, I won’t commit to a date specifically on the backspace issue but we’re well aware in engineering that there are things like that or the default shell for root that needs to be fixed. We’re working on them just as fast as we can and you will see fixes in the “things should just work out of the box” area during this year.

    The first place it will be addressed is is OpenSolaris and distributions that are built from that such as Solaris Express (those distros can be found at http://opensolaris.org/os/downloads/)

  9. […] The Linux developer community is what really keeps Linux steaming along – many vendors and many developers – it’s a train, and it’s very hard to stop a train this big. And this analysis is JUST THE KERNEL. Add in all the libraries, tooling, GUI work, packaging, and other supporting software around the kernel and the community explodes. This is why I laugh when I see “OpenSolaris this or that” from the Sun marketing machine – I took a look and 3 community developers accounted for over 50% of their “community contributions” accepted into the OpenSolaris codebase. That’s for an entire distribution – not just a kernel… I put the community contributions in quotes b/c most of the changes are corrections to comments or minor bug fixes – nothing architecturally significant or evolutionary. OpenSolaris forums are a good Solaris user forums – lots of discussions, users helping users, but very little code. Very similar to Gentoo forums but for Solaris users. Hence even 1.5 yrs later, some changes users have been demanding to make still plague the platform. These issues aren’t new, nor is it just a couple users pushing the idea, but you need developers with architecture (and code commit) control in the community to make anything happen – and generally those developers prefer vendor independence. Anyway, off my soapbox. Someone asked not long ago what my issues were… and that about covers it. […]

  10. […] Earlier this week, I used some data from our traffic logs – courtesy of Statcounter – to make a point regarding the importance of fixing the shell. There are many such points that can be made, but there’s another one today that validates something I’ve believed for some time now but have had difficulty convincing others of. […]

  11. A fix for the solaris ‘backspace’ problem, with the current default shell can be found at this link:

    http://blogs.sun.com/dmick/entry/making_backspace_be_your_default

  12. […] and in a relatively inconvenient location (for me). If nothing else, I need to keep up my “fix the shell” campaign. Unfortunately, JFK had other plans. If I didn’t have project calls this […]

  13. […] that big a deal, but O’Grady was able to show the many search queries that led searchers to a post he did on that very subject. That search data helped make the case that users were unhappy with the shell and wanted […]

Leave a Reply to Nigel Smith Cancel reply

Your email address will not be published. Required fields are marked *