Fear of another GDN/SourceForge

Jun 18, 2006 at 3:13 PM
Just wondering how you guys are planning to "police" the sites and keep things clean and cheery around here. I know it's difficult, but already at this early stage I'm seeing projects creep into that abandoned state.

Take the Office SharePoint Web Part project for example (http://www.codeplex.com/Wiki/View.aspx?ProjectName=SPServerSearchWebPts). It's dormant and about the only thing that has been active on the site since it started is 2 posts, neither of which are from the project owner (whoever it is).

The whole "I'll get to it" mentality shouldn't be tolerated if this site is going to survive. Otherwise you'll have 10 or 20 high-energy, active, thriving projects with hundreds (or thousdands) of one-page-wonders.
Jun 18, 2006 at 10:39 PM
We definitely hear what you're talking about in wanting to keep a good signal to noise ratio.

We fully intend to try to do something, but are still thinking about what approach might work best. Any thoughts there?

Just to give some examples of what we're thinking about, there is the approach of cleaning off the projects that are attracing no interest due to being stale or lacking content. Or we could try implementing a "noise filter" feature that people can use, which when turned on basically will show no content from any project with a popularity of less then X.

Of course both those approaches and other possible ones have their pros/cons, so some feedback on this would be helpful. Obviously the cleanup approach would irk some people who are owners of projects getting "cleaned" and also gets into the touchy game of what is "good enough" to keep on the site.
Jun 20, 2006 at 11:39 PM
As a suggestion, one way to potentially help keep things cleaned up and preventing projects from going 'stale' would be to provide a forum where project starters/developers could announce they are putting their project up for 'adoption' for someone else to take it over. In life there are way too many things that can come up and happen to prevent a very good meaning person from being able to carry on the work that they started. One of the beauties of Open Source projects is that the source is there for anyone who is willing to take it on and continue the work started.

I have noticed within a few projects on Sourceforge that the project starter would make a posting about seeking someone else to take over the work, but unless you already know of the project or drill down to it, you never know that help is being sought. If the forum was at a higher level of visibility then it just might get the attention of some willing developer who is looking for something to tackle to pick up and continue the work on.

Just a thought on how things might be made for the better.

Jun 28, 2006 at 12:33 PM
With the release of CodePlex I think it's important to keep on top of this. Even though you're still queuing up project creation there's a whack of them coming down the pipeline, with a lot of new releases (would be nice to have an RSS on just new releases, no matter what the project is).

It's very onerous for you guys to keep on top of this, so I think some kind of filtering or identification system has to be implemented, or else things are going to spiral out of control.

One thing would be to look at peer reviews or ratings/votes on a project. Have people rate a project in order to provide constructive feedback to the owners. Not only on the quality of the release of whatever it is, but also on the activity and relevance on the site. This combined with your current activity and popularity algorithms should keep things flowing.

Maybe a combination of some or all of these (and more) can make up the noise filter on the site. With enough checks and balances in place, I don't see it turning into a dumping ground if some order is maintained.

Of course that does pose a new question, what happens with mature projects? Let's say I have a project that is done. It's reached version 3.0 or something, there are no new features to add, everyone is using it, so there's no activity on the site (other than the occasional download). I certainly wouldn't want my project to fall into the "we will delete this after 30 days" bucket.

I think there is a lot of brainstorming needed here as there are many great things that can come out of this site, but also a lot of legacy sites (like GDN) that give a stigma to a place like this that you don't want to mimic.
Jun 30, 2006 at 8:38 PM
From bsimsers post:

"Have people rate a project in order to provide constructive feedback to the owners. Not only on the quality of the release of whatever it is, but also on the activity and relevance on the site. This combined with your current activity and popularity algorithms should keep things flowing."

I think popularity and activity provide the indication of quality and relevance. Some other groups subjective idea of relevance (even if it is an extremely large group) is unimportant. If the memebers of the project are active then it is relavent to that population regardless of what the global popularity may be.

As for quality of release there would not be popularity without appropriate release quality. Just as there would not be activity without relevance (not matter how small the group to whom it is relevant).

And any user can provide constructive feedback to a project now by simply making a post in their discussion forum providing any feedback that they wish. If the feedback from the poster is adverse then hopefully they will join the project in order to help remedy that which is adverse.

Jul 5, 2006 at 1:06 PM
I know nobody likes having their work deleted, but I think there needs to be some set of rules around activity on a project.

For example, if the activity is 0 for say 5 days, flag it. If it stays that way for 2 weeks, email the owner. If it's 0 for a month, email the owner that the site will be taken offline (deleted) if they don't act. TANPL1.

Someone needs to come up with reasonable timeframes and state them on the site (basically a service you agree to that CodePlex provides) to keep the dead weight to a minimum here.

P.S. I'm still watching this Office Search Web Part project which has 0 activity since it started. Someone needs to pull the plug or tell them to get their butts in gear and do something with this great resource they have.

1 TANPL - This ain't no parking lot.
Sep 14, 2006 at 3:42 PM
I think this is a really bad approach. There are mature projects that just don't get updated all that often on sf.net. They're still actively used, even if they don't see much development. I don't feel that the level of developer activity necessarily says anything about the quality of the project.
Sep 14, 2006 at 9:33 PM
It's a complex issue, but most of us that use SourceForge (or similar places) can tell you, "We know abandonware when we see it."

It's true that some people might be annoyed when they're told their "Dream project" of creating Quake VII, but some overly enthusiastic developers need a dose of reality.

What I would do:
- 1 month no activity (no downloads or source posted): Warn developers that project will be flagged as "abandonware" in 60 days. Send a "countdown warning" for every 7 days of continued inactivity.
- 3 months no activity: Flag as abandonware and exclude from general search (allow hits if exact project name is entered). Send final warning that project will be completely remove from the system.
- 4 months no activity: Bye bye dustware.

That keeps things fair. Stable projects that are still being downloaded never get flagged. Pipe-dream pages get removed so that you don't go on an accidental graveyard search.

Do that and CodePlex will become FAR more popular than Source-"Boulevard of Broken Dreams"-Forge.

ok, that was my opinion. I now return you to reality.
Sep 15, 2006 at 7:15 AM
Hmm, your comment "We know abandonware when we see it" got me thinking. Maybe it shouldn't be a set of rules judged by a machine, maybe we should give users the ability to vote a project as "abandonware"?

Then if a project is getting abandonware votes and the project owner doesn't voice an objection, it gets cleaned?

Our project delete function gets deployed on September 26, so we'll start cleaning some projects then. Our first target is the projects that were created two or more months ago but nothing has been added or modified yet (even the Wiki homepage). There's no data loss since nothing was added, so if the project owner comes back later we can just re-create it for them.
Sep 15, 2006 at 4:56 PM
DrPizza has a very good point... a mature project might not have much development, but be benefitting a large number of people.

Perhaps a combination of measures would be instructive to gauge how much "quality" a project has. These could be the traditional ones (developer input) as well as user community feedback, something like an Amazon review system with stars and all that. A five-star product is still a winner even if it isn't updated regularly, or even being produced anymore.
Sep 15, 2006 at 8:18 PM
Far more pressing an issue than abandoned projects are those that never get started. While eagerly read the RSS feed several times a day, the majority of the projects that come up don't even make it to the point getting a project description, let alone posting code. As I type this, of the 5 most recent projects, only 1 has a description, and none have posted source code. I am nearing the point of not even bothering to check out the new "projects" anymore, as they are mostly vaporware.

Essentially, a comittment, or at the very least intent, should be demonstrated before projects are made public. While source code should be could be a requirement, just requiring a project description improve the situation.

Is it so unreasonable to ask the initiator to describe what their SlicedBread Managed Framework actually is?
Sep 15, 2006 at 10:34 PM
I agree with that. It seems to me that instead of announcing new projects on creation they could be announced on their first release (personally I'd go so far as to say that they should (a) release something (b) modify their homepage). This would at least require some output from them.
Sep 25, 2006 at 3:10 AM
Maybe some enhanced self-policing.

Maybe the project owner needs to log on and tick a "retain this project" box if there is no detectable activity for say three months.

He gets an e-mail etc. but he has to do something.
Sep 25, 2006 at 3:14 AM
The danger of external reviews is that a project might be very useful to a small group, while the majority of "judges" can't even see why it exists. The "judges" vote it down and actively mess up the lives of other people.
Sep 25, 2006 at 4:49 AM
I really like the idea of voting for projects, much like a digg type environment. I also like the concept of expiring projects if there's inactivity (and the schedule someone proposed seems like a good framework to start). Could someone turn this into a work item for a future version of CodePlex?
Oct 26, 2006 at 5:29 AM
MikeGale -

Only in malicious situations would external judges vote it down if there are positive uses of a project.

Plus, the idea was, as I understood it, not so much "voting" a project off, but by combining a number of metrics to determine the value of a project. Maturity would be one such, so the mature projects would get promoted to a higher generation and not get garbage collected, even if there were no immediately visible references.