intertwingly

But Sam, what did you think about it?

And if you have some time this evening, come to our regular Thursday evening weblog writers meeting at Berkman.

http://blogs.law.harvard.edu/thursdays

It would be great to have you be part of the discussion.

Yeah Sam, what did you think?
Inquiring minds want to know

Dave, the keynote covered a lot of ground.  And I have a lot of thoughts.  The two biggest areas where I have thoughts is that I do believe it is possible to make a decent living writing software.  And I do feel that RSS needs a profile.

On the latter subject, I feel that Charlie is right: the task needs a moderator who's not self-interested.  And while I understand the motivation to keep the discussion group small, any group which excludes me is the wrong answer.  ;-)

On the former subject, I agree that it is hard to break into the shrink-wrap market.  I go back to your analogy: I pay $1 to ride the subway downtown. It costs $300 to fly to NY and back (two hours in the air). A cab ride to the airport -- $40.  In other words, you can work for the MTBA, or figure out how to run a profitable airline, or sell services to individuals.  If you want to sell cars, you have to convince people that your cars won't break down, and when they do, they will be able to find parts.

I will try to find the Berkman center tonight.

I wasn't at OSCOM to hear the keynote, but this reads like a bunch of kids fighting over control of the sandbox. How accurate is my assessment? Or were the arguments in jest?

Please don't get mad at me, I'm just interpreting in the third person and want more feedback.

comments on Sam Ruby's blog ...

Randy it was very high level discussion as far as I was concerned. I was surprised how much Sam agreed with me, he was sitting in the front row so I could keep my eye on him. ;->

I don't have a problem with Charlie getting in the loop here. To be honest, I've been trying to do that since I met him. We need users in this space to start making their presence felt in a strong way.

In any case we didn't solve any problems, I put in a pitch, many times, for people working together. I stated many opinions, and asked for many opinions, and for the most part it was like a very good mail list, and then it kind of flamed out.

Onward. YMMV. IANAL, etc.

Dave, do you agree with need IETF to settle the multiple camp RSS and BlogAPI thing?

Sorry guys, my sloppiness again, I meant "do you agree we need"

Dave - how can I not agree with you when you say that open source and commercial companies should work together?  ;-)

After all, I work for a commercial company.  I also am deeply involved with one of the most respected open source organizations that exist.  I don't merely believe that these two worlds can coexists, I actively work to make it happen in whatever small sphere of influence I happen to have.

Ooof. Reads like it was quite a keynote....

Today is the second day at OSCOM 3. It opened with the keynote by Dave Winer. I am still digesting it. *** Coming back to my earlier post, today it looked like we are still far away from interop in the blogging world. *** UPDATE After reading other...

"in whatever small sphere of influence I happen to have" ... Sam Ruby? Small sphere of influence? That'll be the day ;-).

Personally, I don't give a crap if the software is open source or commercial.  If it solves a customers problem, use it.  End of debate.

Software doesn't have to be bug free, commercial, open source, created by 40 people or 4000 people or whatever to be good software.  It just needs to solve immediate needs.  Anything beyond that is a waste of time.

James, solving immediate needs is cool.  However, I add a criteria that I don't want software that leaves me stranded.

Generally that means that at least one of the following is true:

• the software has no bugs of any importance to me, or
• the software is well supported, or
• I have the ability hire somebody to make the changes I need (possibly including hiring myself).

Acceptable criteria.  The ongoing debate, however, seems to fail to account for the fact that it simply does not matter if code is open or closed source, written by a large company or a single individual, or whatever.  Using "good code" (see definition below) is a standard of good software development practices: using bad or unsupported code cannot be seen as "solving the customers problems" as it just creates new ones in their place.

Now, if someone creates code that they intend other people to use, they should be willing to accept a certain amount of responsibility: either support it or open the source so others can support it and get out of their way.  For me, this is the key attribute of "good code".

My bottom line, however, is that debates over which is the best model (closed or open source) is a worthless waste of time.  Just write code, take responsibility for it and move on.

Why don't we all just support Mark's IETF draft and get this over with. http://www.ietf.org/internet-drafts/draft-nottingham-rss2-00.txt

James, so there is something "beyond solving immediate needs" that is not a waste of time?

To me, moralistic terms like "good code" are subjective and therefore not particularly helpful.

Not all open source software is "good".  Not all commercial software is "bad".  But this admittedly secondary attribute can be very important - if only to the extent that it may give me confidence that I won't be left stranded.

Is it a reliable predictor?  Absolutely not.  SourceForge is littered with abandoned projects.  And some commercial companies provide excellent support.  But is it an important characteristic?  Absolutely.

Understanding what software people are willing to put up with, and why, is something that interests me.  You are welcome to participate in the discussion, you are welcome to ignore it, but it is something that I chose to waste my time on.

Sam, you have hit upon exactly why open source is so important.  Once you realize that "not getting stranded" is an important criterion, open source is the ONLY safe solution in the long run.  Because:

- you may discover that a product that appears to have no important bugs actually does;
- a product which is well-supported now may become poorly supported or unsupported over time;
- BUT if the product is open source, you can ALWAYS hire someone (including yourself) to make the changes you need.

In my day job, we recently ran into a problem while indexing PDF documents -- our PDF-to-text library was "obeying" a bit in the PDF documents that prohibits selecting/copying text in Acrobat Reader, and interpreting that bit by refusing to give us access to the raw text of the document to index it.

Obviously, a PDF indexer that doesn't index PDFs is not terribly useful, and does not "solve immediate customer needs".  And naturally, because this only affected 8 of the 5000 documents we had, we only discovered the problem a week before pressing CDs.

Luckily, the library in question was open source (part of Xpdf), and I was able to download the source code, comment out the "if bozo bit is set" line, recompile, and index our last 8 documents.

I will NEVER, EVER, EVER bet my business on software to which I can not get ALL the source code, on demand, the week before an immovable deadline.

Mark:I will NEVER, EVER, EVER bet my business on software to which I can not get ALL the source code, on demand, the week before an immovable deadline.Via ...

Pingback from Bitflux Blog :: NEVER, EVER, EVER

Mark:I will NEVER, EVER, EVER bet my business on software to which I can not get ALL the source code, on demand, the week before an immovable deadline.Via Sam ...

Randy, I don't necessarily support registering RSS at the IETF, or not. I have mixed feelings. I don't really get what the IETF is. Whenever I say I don't like something the IETF did, people tell me that isn't what the IETF does.

There really isn't a standards organization that can handle this. For the reasons Sam outlines, the W3C ruled itself out by taking sides in the RDF fork. They're an interested party.

I would be interested in seeing some serious users take a leadership position here. That might be the only way to get unity. Developers care about things that are, to be blunt, esoteric.

My guess is that RSS is never going to be more unified than it is now, nor do I think it has to be more unified, it would just be nice if it were.

Thanks Dave, for listening. Would you ever consider the RSS 1.0 fork or is their too many hard feelings? I agree 2.0 is better, but for me, I'd take a bad constant over a good variable.

Does anybody know if Mark Nottingham is moving his initiative or sitting on it or given up?

I got a good reply for Dave Winer on his thoughts about the 2.0 draft-standard. I also posted some thoughts on the Dude blog. ...

Aparenty I missed some great keynote by Dave Winer. Sam has some bits about it as well, and I feel reconfirmed in arguing (link lost due to blogware change) that CSS vendors and the OSS community need to get together and _talk_. I think this is...

Randy, I assume you're asking about UserLand, there is a continuous path from 0.91 to 0.92 to 2.0, where each new version added features that are optional, so you could upgrade just by changing the version number and add features as time permits. 1.0 took things out of the format. Now I admit I never thought RDF had a place in RSS, but even if I thought it did, there was never a way to do the corner-turn they wanted us to take. It's kind of why Apache is having a hard time (I hear) getting people to use the new HTTP server, it doesn't run all the old modules. We had applications, deployed, for all the stuff that they didn't want.

re: "all the stuff they didn't want"

...
2002-04-06: Apache 2.0.35 released
2002-06-18: Apache 1.3.26 released
2002-06-18: Apache 2.0.39 released
2002-08-08: Apache 2.0.40 released
2002-09-24: Apache 2.0.42 released
2002-10-03: Apache 1.3.27 released
...

What we are witnessing is a multi-year corner turn.  Both releases are actively maintained.  Nobody is being left behind.

Here's an Overview of new features in Apache 2.0

Sam, I wonder if the users feel the same way. When you say they both are being actively maintained, are new features going into the legacy version? What if the user wrote their own modules, do they have to transition those? What if the programmer is long-gone? How was the decision arrived at that compatibility wasn't important?

BTW, these are questions. I don't know the answers. I'm not arguing. Just wondering. Given past discussions this kind of disclaimer is necessary.

I've seen a number of such corner turns in the Open Source world, and the answer is most decidedly, yes... new features can be, and often are, added to both branches in the evolutionary tree.  This may go on for months or years.

What generally happens, however, is that a Darwinian selection ultimately emerges as everybody coalesces around one version - not necessarily the one that some elite had chosen, but the one that users want.

One thing that is always true, however, is that you are never stranded.  Even if you find yourself to be the only user who has chosen a given path.

For these users, there is no trunk.

> For these users, there is no trunk.

It'd be good to note that there are still users of NCSA httpd -- the
web-server that Apache replaced.  They've even had a release or two
since Apache became popular.

Lauren Wood:  I feel the same way. I'm on my third iteration of weblogging software and I left the first two because they violated the third rule in my opinion at the time, namely It looks like people still work... [more]

Trackback from Sam Ruby