Back to Questions
2.3k
views11
comments
Going live?
f#

This is just a general curiosity question. When do you suppose this

site will be going live and viewable to the public? Next month, a few

months, a year? Just curious that's all.

.

That's a good question. I think a few things have to happen to go-live.

  • We need a lot more questions on the forums, especially language questions. We especially need a bunch of "dumb" questions like "how do I loop", "how do I do recursive stuff", "how do I use unions", "why do upcast and downcast work like they do", etc. We need to build beginner questions and how-tos. We know that the F# documentation is a good starting point, but it is lacking and could use a hand from what we all learn here. I personally read the F# source code for questions now (I must think learn to think more like Don), but I doubt that many beginners here will feel that is a good way to approach a question.
  • We need a public-facing General Discussion forum. I'll create one of these, but I don't want to do that until we get more activity on THCF. I don't want the place to look confusing with two "how to forums" and have contributors wondering where to ask a question.
  • We need articles (I have stuff I want to co-author, just haven't found time and co-authors). I'm talking to people on IM, but things are moving slowly.
  • We need a few cool projects that demo F# (I have some of that as well, but need to get people involved). Syme and Margetson got the ball rolling. I want to make some momentum on that. My personal competitive nature finds me in a spot that I am compelled to make a few *more better* demos than those guys (I got street-cred to make good on).
  • We need coverage of F# on mono/SSCLI. I need to re-address this issue, i.e. I need to get up to speed on Solaris with SSCLI/Rotor/Mono and F#.
  • We need a reasonable status on ReflectF.

We just got our source repository running and I haven't gotten a chance to push any code up there yet.

Our original goal was to go-live on April 8th, 2006. Many members have busy schedules and haven't had a chance to install F#, participate, etc. Some members may not be able participate. Travis Feirtag's wife just had a baby; he'll be out for a while. Scott Isaacs wife is due in six months and three of the local guys (grh, damon, sisaacs) and jhaley are all moving to new houses within the next 90 days. Plus we all have jobs. I tried to make a wide enough base group so that even during busy times, the flow on The Hub would be good.

My estimate is likely to be a few months, but we'll see how things progress. I'm spread thin, with development efforts, F# "tutoring", my growth in my own personal language competence and ReflectF. But I'm very comfortable with what has started. I think as people engage on a more regular basis and start to actually use and get comfortable with F#, we'll see a lot more activity.

So, one month is likely too soon; one year, too long. I also hate to give an estimate because things may change in tempo. But I'd guess at this point, unless there is a drastic issue, and I personally can't get the initiatives in place and moving, that another 3 months (June-July) would be a reasonable estimate.

We don't want to be "elite", but rather competent. We will do the F# language a better service with a longer duration incubation that results in higher quality. Unless I'm mistaken, there is not a compelling reason to expedite at this point and I'd rather have members feel they have freedom to learn and contribute, rather than a more mandatory or demanding stance on my part.

I'm glad that you asked. Others may want to understand these motiviations and this is a good place to discuss it further. I like what we are doing. I think that these early steps to engage and go are the most difficult, but also, the most memorable.

---O

p.s. How are you doing with F#? Anyting that I can help you do?

By on 3/12/2006 1:42 AM ()

I must say I'm really keen to see the hub go live. The F# List is not a great forum - the list server sucks, and the hub is SOOO much better and more exciting to be a part of. I feel you would see a substantial number of the F# List participants hitting the hub as soon as you go live. I will gladly write a strong referral email on the F# List, suggesting that the hub is a great place for beginners and experts alike.

The only things I dislike are that I find myself having to look in multiple places to see what's going on (www.hubfs.net, cs.hubfs.net forums, blogs, individual blogs etc.) But I don't really know how to improve it.

Of your "must have" list I think you only need (a) source management and (b) articles.

In other words I would argue for a go-live date in mid-April, using the dealine to squeeze articles out of myself and others.

By on 3/23/2006 4:04 PM ()

Ugh. I wanted to avoid more discussion on this just yet. OK.

Plans:

  1. Move hubFS.net to a new hosting provider. Oochie provides great service, but we will outgrow their bandwidth given what I know. This is not critical for go-live, but I'd like to move before go-live.
  2. Source code repositories are up. We have started basic projects. Swarer is writing a blog post to cross post Source Control for projects.
  3. We need articles. I have to make a choice here. I have some stuff to release (and have been a lazy bum in getting it out; actually, have been busy behind the scenes), but I've been focused on a number of things and haven't given this time (to complete). I'll raise my personal priority here.
  4. We are tossing the DotNetNuke side of The Hub. It serves no purpose at this point. Given the direction of CS, I'll wait and grow with them on a unified platform. The primary functionality of DNN was schedules/events and that was low priority.
  5. I'll open a "Public General Discussions Forum" for anonymous users as the start discussions here point.
  6. Deeper in Dot Net 2006 is April 22nd, 2006. I'm concerned about logistics and time, given the move and my dependence on a few members of WI .NET Users Group for hubFS maintenance. April is definitely within reach, though.
  7. We do have a "community license" from Telligent which is the equivalent of their $8999 enterprise server product, i.e. we'll scale well. Once moved, we'll install that license.
  8. We need one killer app. I've taken jjbHop "underground", i.e I'll be doing that on my own to get it done. I want to retain rights under jjbr for the moment, but will associate the project with The Hub as well. It will be pure F#. ReflectF needs attention as another project. I have to start communicating more, but the source is up on the respositories (swarer will give details). This doesn't need to coincide with go-live, but it would be useful. Hop has a number of features that will be useful to the general public (anyone who uses a browser) and I have long term plans for it with jjbr and IronWake.

Observations:

  • The first round of people on The Hub served their purpose. I found those that are serious in commitment and those that are not. I had hoped for a few more, but we are finding more people that are interested and will be able to contribute. Current membership will stay open for those that are not active. Go-live may inspire some of them to get over here and make stuff happen.
  • Source management was planned for THCs exclusively. I'd like thoughts on this. Software releases would be rolled up with a corresponding article and a zip or msi would be made available as an attachment. We have the Software Releases and Article Releases on The Bank.
  • My biggest fear is that we open, don't look busy and the "collection of community" fizzles. There is risk in that no matter what approach we choose. I just want to be informed on the go-live decision. I don't want to be "that F# site that I saw and nothing was happening".
  • Don is right. The F# list server is not "friendly". Getting The Hub running as a gathering point will be useful. Pretty interfaces make for repeat visitors. Sooner is better on this point.
  • Regarding a "spread out" feeling, recall that the outside world will see only what we post to The Bank. Any THC can post to any of those blogs at will. I have no editorial control in place at the moment, but will likely put in moderation, at least initially, i.e. I or someone else will have final say on releases on the public blogs. I want to control image, both initially and in the long term.
  • Regarding a "spread out" feeling on the forums. That is by design. If you aren't familiar, use the "posts you haven't read" link in the right side-bar. This will show you every post (either all or in your filter) that you haven't seen. Get comfortable with filters and your inbox as methods for getting around and reading.

I like what we are doing. This is timing now and feedback is a good thing.

Tell me more . . .

((Andy's innocent question gathers no moss))

By on 3/23/2006 7:42 PM ()

> We'll push to go-live (I need a few days for logistics, but there

> is a real chance that we can go-live next week). Let me

> get organized. We may take The Hub down for 8 hours to move.

I think a rollout of cs.hubfs.net in the next few weeks would work. Here's my suggestion for a "go-live" plan:

- ensure the thing is secure

- improve the front page (link is to a separate discussion on this)

- choose 5 initial forums (link is to a separate discussion on this)

- move as much existing discussion content as possible into public-facing forums. Suggest you only have two private forums.

- Ensure 3+ good articles. You can cross-syndicate my First Class Events article. I'll write another on "operators in F#", to go on the hub. (One reason it's good to go live is that people like me can write an article explicitly for the hub and cross-reference it from my blog. Without going live we continue to write only for our existing blogs. Writing for the hub occasionally will lead blog readers to the hub and let them see that it's "OK" to write articles for the hub)

- I pre-announce to the F# list and on my blog that cs.hubFS.net exists and that I see sites like this augmenting and eventually subsuming the F# List. You follow up with an email that issues an open invitation to write articles and cross-syndicate existing content, offers to permit material to be cross-syndicated on external sites, and a note that (briefly) explains the purpose of hubFS.net its purpose and saying it will go-live around a particular date.

- have one example of cross-syndication of external material onto hubfs.net (my First-Class-Events article may be a good example).

- go-live itself can be low key if you have concerns. It can even be unannounced, to let google etc. index. When things are satisfactory, market as you see fit. I would avoid a huge fanfare or over-selling - this thing will attract attention by consistently hosting good content over 1 year, serving a vibrant role in forums, and through high page rank. You will get lots of people looking, a few initial lively discussions, a hiatus, then people gradually returning, then a consistent and growing level of activity

- Very important: drip feed after launch. Before you go live, have prepared a "drip-feed" of material for 1-2 months. Have a few good articles ready. Have a few blog entries you plan to write.

- Don't try to seed discussions in the forums - let users initiate and gradually feel like they 'own' the forums.

All very obvious, but I hope helpful. If you like use any of the above for a "go-live" plan for cs.hubfs.net.

And you're totally right - let's have fun :-)

By on 3/25/2006 4:45 AM ()

Given that things will "heat up" around Deeper in Dot NET 2006 between now and April 22nd, I'll guess that sooner is better. I'm honestly thinking between now and March 31st, 2006 for go-live.

I also have a 3/31/2006 Milestone for IronWake that I cannot miss. Guess it will be a busy week :-)

Logistics:

  1. Complete logistics plan. The Hub is secure. The source repository will have issues in the near term. We can "lock that down" temporarily until we adjust this. I don't have time or resources for a quick forms authentication app for password changing for swarer's stuff. That is what is needed. Other hubFS members will contribute that, but I cannot guarantee status.
  2. Sign up with target ISP.
  3. Reposition site on "test domain". Perform dry run of move of cs.hubFS.net. Upon completion of dry run, announce a planned shutdown for the move. The move should take a total of 8 or less hours. Test CS redirectes for www.hubFS.net and cs.hubfs.net with the goal that www.hubFS.net may be once again replaced with a portal that enables other features. That is not the immediate goal, but the options should be available in the future with minimal effort.

Post-move changes:

  1. Front Page will change. I'm old-fashioned, so I've ordered an eight foot heavy timber doorknocker.
  2. I wanted a lot more articles. I'll get my butt in gear and get the rest of my materials together. The Drip Feed is the exact issue that I want to address. There should be enough content for periods when The Hub is slow and there is little activity.
  3. I'll have two articles at a minimum. I'll also co-author one with Scott Isaacs. Chad Albrecht is working on a grid computing with F# article and tending to details on that. Mike Jones' article needs an editorial swipe, but is useful.

Re forums: I completely agree. If people come, they'll find their topics and push the discussions as they see fit.

I agree on announcing as well. Let the site speak for itself. I wanted a bit of fanfare on opening, but a few blog posts from members (and a link on the F# mailing list) to direct "first round traffic" will probably suffice. It is better to earn the respect and trust of the F# community, rather than gratuitous linkage.

I'm organizationally-challenged, so the forum reduction will be traumatic. Give me a day or two to bid farewell to these old friends (that have been chastised for their lack of content, for which they are not responsible; they waited innocently and patiently to be filled).

By on 3/25/2006 5:35 AM ()
  • :-)
  • I've not run a web server used by more than 2 people, so I don't know about machine requirements etc. It sounds like the limit re source-code management is "hub tutorial projects" - is that right?
  • I think you're getting the balance right - you're creating a great site
  • I don't think you need a "killer app". Don't worryat all about that - people will come here to learn F# and be part of community, not to be awed. Anyway we have a proven killer app - interactive mass data exploration and analysis, of the kind I described in my one-and-only hub blog entry. You couldn't have done what is described there in the given time and LOC in any other existing language implementation, due to the DB connectivity requirements and the perf requrirements.
  • I think quite a number of existing posts can be put into the open-for-external-view folders, e.g. my one on trees etc. was written with this in mind. That will help populate the forums.
  • The number of forums just has to be reduced. It's the only negative thing James said about the site when we were at the pub tonight :-)
  • I'm off to bed now and will be not online much over the weekend!
By on 3/24/2006 7:12 PM ()

It's going to be interesting to see what happens when it goes live. I have already seen F# gaining interest in some areas with some folks I had never thought would have been interested in it. For instance in a blog I read by a lady who is a hard core C++ and sometimes Java person she was wishing for a language like the Scheme she used in school for doing generic algorithms. Several of her readers recommended F# before I could even hit the submit button on my response. So I know if it is gathering interest in those circles it is surely gathering interest in other areas as well. These are folks like myself who are not generally Windows people at all either so you know if it actually attracts types like myself to install Windows and Visual Studio and .Net it must be something that people have been wanting for a while. That and the whole fact that it has a great debugger in the IDE of choice for it which is something Lisp and Scheme have never had.

On a side note after I installed Visual Studio 2005 Pro and ran the F# installer thing everything worked right off the bat. I had zero setup problems and that never happens with other platform IDE's except Eclipse.

By on 3/24/2006 2:16 PM ()

OS asked me some questions in a private email regarding what sort of projects might be hosted, and whether I thought there was a need to actively host student projects. basically I don't think so - I was more thinking that enthusiastic students like Dagur should be given the opportunity to host the tutorial material they are presenting on the hub.

I don't want to push you into hosting source code for really active projects - you can use sourceforge etc.for that, but I feel you should angle to have the hub as the home for tutorial projects. Also as a portal for the major projects that flesh out F# with visualization, numerical etc. support.

> 1. Is your expectation that the repositories will be

> public, password protected or otherwise?

Can we leave the public/password-protected up to the individual project?

> 2. Do you expect that The Hub will take on each student as a member?

I would hope to see lots of students on the hub. But I don't feel students should be a special category - some students will be community leaders, some will be beginners, some will be incredibly annoying by ranting in ways they shouldn't :-) The same criteria should apply as to everyone else - e.g. community leaders get invited to contribute articles, or syndicate their existing content through the hub.

> By default, students will see all public stuff, i.e. public

> forums as described and "The Bank" blogs. I don't know what

> is appropriate, i.e. full integration of the groups, or

> separation for differing conditions.

I would like to see all the existing blogs made public (including our F# Core Team blog). I'd say the only non-public thing should be two hubFS contributor/administrator "explicitly private" forums. And even then existing posts in those directories that are clearly "OK" for general consumption should be redirected into a public-facing forum.

> Do you have an estimate for traffic and file size for

> the source repositories as needed for the student population?

> Dan has space, but I want to size this correctly up front.

I'm thinking you could host good, actively developing tutorial and min-project material for consumption by the whole community, including students. Some of this will be projects performed by students.

By on 3/25/2006 3:53 AM ()

Excellent. That clears up the most difficult issues.

Re 1: Password protected by default will be the standard. I'll let swarer decide policy on public consumption.

Re 2: Students are a "special" category, aren't they? ;) Seriously, all hubFS contributors have rights to blogs public or private. So, if a student is anonymous (under CS login standards, i.e. not logged in), then they can post all they want in the forums. But if they become a contributor, then they get the privileges of contributor status, i.e. a blog, article publishing privileges, source repository authoring access, etc.

Re blogs: Ugh. The pain is overwhelming. Not . . . in . . . my . . . plan. Must . . . be . . . flexible. The Microsoft MVP program has "the coffeeshop" for all behind the scenes communications amongst that developer community. We need a private place for that stuff. I wanted control over the content that is released. I was considering TheServerSide.NET as a model. If the proposal is to "just open stuff up" on blogs, then anything that anyone posts will make the front page news. ((contemplating)). Also, I suppose that means that we allow anonymous commenting on blog posts, too, or allow authors to decide on a per blog or per post basis.

Agreed on content though, i.e. any private forum topics that are appropriate for general consumption should be moved to the public forums as soon as possible.

I must admit a few things here:

  • I'll be relieved when we open our doors.
  • Organic growth has advantages, especially now that most of the tough issues have been resolved w.r.t. operational/administrative issues.
  • I'll feel much less pressure to produce on many of my topics as the community will "just produce" what it wants to produce.
  • The initial invitee list is/was useful. We can decide on status of each individual at a later date.

We do need an anti-spam strategy though.

Ok, I'll relax a bit. Give me a day to digest these thoughts and to make sure that I don't take any liability in these policies.

By on 3/25/2006 4:45 AM ()

OK - I understand a bit more (re "wanting control of content posted", apart from the forums). The problem is that if you give us blogs, we start writing as if the blogs were public facing. e.g. when I write an announcement "F# Version 1.87.3.1 released", I expect everyone to see it - and if not, I ask myself "what's the point of a blog that can only be seen by <1,000,000 people"? :-)

Perhaps you can solve this by making the blog entries much less prominent on the front page, i.e. instead use the featured article/forum-discussion/blog-entry model to give you editorial control. You should certainly exercise strong editorial control on this site.

You could maybe also simply not give so many people "contributor" status - instead just post people's articles. Or else not give contributors a blog by default.

By on 3/25/2006 4:57 AM ()

Right. In the end, it comes down to trust. If I invite a person to participate as a contributor, then that person has received a level of trust from me or whatever administrator brought them in. At some point, I have to trust that people will act correctly.

There are "advanced options" that allow a per blog or per post entry to be aggregated to the front page (site) or aggregated to the group (THCB or The Bank).

By default, the post shows up on the blog RSS for the user.

I'm trying to determine what would be a beneficial policy. By default, syndicate to the front page, or by default, don't syndicate to the front page. ((thinking))

The old model was that anyone could post to the "Articles Blog" and determine their level of syndication and allowing moderators/editors final say on the post to the front page.

THC Blogs will go public facing by default. I just have to consider some rules here. I'm trying to decide if The Bank has the utility function that I intended in light of this new approach. I think it does as a preaggregated view, but I want to cut waste now while it seems useful.

By on 3/25/2006 5:09 AM ()

    Topic tags

    IntelliFactory Offices Copyright (c) 2011-2012 IntelliFactory. All rights reserved.
    Home | Products | Consulting | Trainings | Blogs | Jobs | Contact Us | Terms of Use | Privacy Policy | Cookie Policy     		Built with WebSharper