Tuesday, February 24, 2009

Features, Requirmements, Defects, and Enhancements

Features represent what a product can do.
A feature is composed of a series of requirements.
A defect can be associated with a feature.
An enhancement can be proposed for a feature, but it is not a requirment.
An enhancement can become a requirement (or a requirement can be created from an enhancement request.)

A feature is implemented by a series of tasks.

A defect is not an impediment, but an impediment to feature approval can be tied to a defect.
Approval is a task that indicates a feature is complete (or at least acceptable in this iteration.)

Tasks & Impediments

Tasks represent work to be done.
A task can belong to a category (such as development, qa, system administration, etc.)
A task can have a work estimate (such as five hours)
A task can have subtasks (groupings and dependencies?)
A task can have comments (which can be about planning, implementation, resolution, etc.)
A task can have a status (such as not started, in progress, complete, blocked)
A task can be blocked by an impediment.

An impediment is something that prevents a task from being done.
If one task depends on another task, the predecessor is not an impediment -- it's being incomplete is the impediment.

Friday, January 30, 2009

Blogs, and Emails, and Projects

Okay, I should really consolidate blogs. I keep saying that.

I just posted over at http://one-shore.blogspot.com/
What do you want in a software test consultant?

For the record, this one is oneshore.blogspot.com and is my original one-shore blog. I also created another at wordpress oneshore.wordpress.com when I couldn't get into blogger. Actually, I think it was my fijiaaron blog that moved over there (used to be here ). Fijiaaron started out as my travel blog, when (not surprisingly) I was in Fiji. I also sometimes keep blogs of projects I'm working on such as qa-site and fluffy. To make matters more confusing, the qa-site blog hosted at wordpress and the fluffy blog is hosted on my own install of wordpress on it's qa-site.

Actually, it started as my web email address when I started traveling, at fast-mail.org. I still use it, Abut I also use my one-shore email as well. And keep klamathsystems.com as my main point of contact for recruiters, but they've drifted into my gmail as well. So I have 4 email addresses that I monitor.

I also have stuff over at herculeangrp.org, and projects for Harvest Requirements and ForgetMeNot.

Some discipline is in order. And some focus.

But I also like to cast my net wide as well. I do testing, and some systems work as need, and web development (both back-end stuff in Java/LAMPPPR), and the HTML/Javascript/CSS stuff) . So when people say "what do you do?" I say "computers" which I think most people translate into "graphic designer" which I'm decidedly not.

Monday, October 27, 2008

Distributed software development collaboration

I'm reading a book, "The World is Flat" by Thomas L. Friedman. It's a pretty good read, with surprisingly informed descriptions, and fairly accurate diagnoses.

The main thrust of the book is how technology is changing the world (making it flat), and it talks about offshoring and global competition. It's definitely food for thought for me, since I'm thinking a lot about that. The "one shore" idea is that you can have a local or distributed software development team, and increase or at least maintain productivity, using new technologies and tools, particularly open source tools.

I'm not trying to build an offshore consultancy, except in the sense that I want to be "offshore", literally, and still working from my sailboat. But I'm not opposed to collaborating with offshore individuals or teams.

As a matter of fact, two strategies I'm interested in pursuing are 1) collaborating with other independent developers, testers, sysadmins, designers, project managers, etc. and 2) building a team from local talent where I'm living, whether that be Seattle, Ecuador, Fiji, or wherever.

The products and services that I'm targeting are software development lifecycle and collaboration tools.

For the first form of collaboration, the idea is to enable a distributed team to work as efficiently as if they were centrally located. Not, something I believe completely possible, but a goat to strive for. A lot can be done in this less than ideal collaboration environment, and there are definite advantages, including the real productivity gains that can be achieved by those that prefer this environment, and the enabling it does by spreading development over a wider geographic pool. Not to mention the typical aims of traditional offshoring that include reduced labor and infrastructure costs.

For the second for of collaboration, the idea is to develop best practices and tools for teams that might otherwise not be able to pool the capital and experience that large scale professional teams in big budget organizations can muster.

The primary purpose of both forms of collaboration for One Shore, and for me specifically, is not, as I said, to build an outsourcing consultancy, but to develop the tools and processes for my own development aims. So the result is the same whether it's a widely distributed development team building the tools that increase our own productivity, and learning the techniques to work disperse, or whether it's building the tools that allows a team of talented, but untrained individuals to work together in a way that typically, only experienced enterprise groups can collaborate with big budget software.

Monday, June 30, 2008

Whoops

I just created one-shore.blogspot.com thinking that this blog had been deleted. Basically, I wanted to say what I'm working on (in my spare time) these days:

One Shore -- my company website for consulting (not actively working on it)
QA Site -- my project for building hosted tools for testing
CuencaTravel -- a portal site based on Joomla started in Ecuador (currently on hold)
PM Site - a spin off of qa-site.com geared towards tasks and requirements instead of bugs and tests

And my new project, code named "fluffy", researching Flex testing tools and strategies. I'm working with Nate on this. He's kind of mentoring my learning Flex. We brainstormed the
idea at a BBQ last Monday. I set up a repository and QA site with all the regular tools: bugzilla, projectpier, trac, blogs, and forum.

So far, I've:
I'm pretty impressed with FlexBuilder. It seems well put together. It's impressive for an Eclipse plugin, and I did a lot of work with custom plugins at Varolii. FlexUnit seems to be a half-measure, but probably gets the job done, as far as testing ActionScript at the component level. ASUnit is another project (open source) in the same area.

Friday, February 15, 2008

On wordpress

I've been doing most of my posts on Wordpress lately at http://fijiaaron.wordpress.com and http://oneshore.wordpress.com. Because I like the theme there and I have built up quite an archive. I'm going to resolve to give Blogger another shot.

Like I mentioned on my last post there, At least it has a bigger text area. But adding images is a real pain on blogger. Pointing at custom URL like http://blog.cuencatravel.com is easier with blogger, but hosting my own is easier with wordpress.

Saturday, January 5, 2008

Requirements as versioned documents

A requirements document is almost always under revision control. Any requirements management process worth it's salt has a change process. But individual requirements as versioned documents... Again, requirements can be tied to code releases. And it should make the change process easier, as well as querying for requirements. A requirements compatibility matrix could be generated (based on versions of each requirement.) Requirements implemented in this release is just another generated document, based on the source control logs.

Tests as versioned documents

A test under version control could also be branched with a code release. A versioned test suite could change to meet new requirements. A tag of tests could be a suite. Versions could even keep track of test completion status and actual results, but that might be going too far.

Bugs as versioned documents

A bug that is a document under version control could have the heading and description edited without worrying about affecting credibility. Previous versions of the bug would contain the original text. A bug under version control could be branched, and branches could be tied to certain code branches or tags.

Document centrism

A test is a document
A bug is a document
A requirement is a document
A task is a document

A document should have the following capabilities:
Editing,
Versioning,
Commenting,
Sharing,
Searching

A semantic document (like XML) could also be queryable, and have better searching. It could be populated from a template plus data.

Reports are documents generated from a template plus data.
Forms are applications that retrieve data, potentially by processing documents.

A collection of tests, requirements, or bugs is a report generated by querying multiple documents.

System Integrator

According to wikipedia, One Shore pretty much matches the definition of a system integrator.
We pretty much build a QA system from open source parts. A QA site is a hosting solution, a Linux server or VPS, a collection of open source tools, including version control (svn,cvs), build automation (ant, maven, luntbuild), bug tracking (bugzilla, trac, mantis), test management (qatraq, test-site, testopia), automated tests(fit, testng, junit, selenium, watir, samie)project management (dotproject, phprojekt, trac), document management (knowledgetree, alfresco), wiki (dokuwiki, dekiwiki, moinmoin, trac), forum(punbb), blogs(wordpress-mu), etc. and integrating them.

I noticed trac is in there three times. Fit and trac are popular for a reason, and I think it has to do with document-centrism.

Friday, January 4, 2008

Java EE5 setup

Installation Directory [/opt/SDK]: /opt/java/ee5/sdk
Admin User [admin] : admin
Admin Password : secret123
Admin Port: 4848
HTTP Port: 8080
HTTPS Port: 8181
The following items for the product Java Platform, Enterprise Edition 5 SDK
will be installed:

Product: Java Platform, Enterprise Edition 5 SDK
Location: /opt/java/ee5/sdk
Space Required: 215.65 MB
------------------------------------------------
Java 2 SDK, Standard Edition 5.0
Sun Java System Message Queue 4.0
Sun Java System Application Server Platform Edition 9 Update 1
Sample Applications
Java BluePrints
First Cup of Java Tutorial
Open ESB Starter Kit with JBI and BPEL

------------------------------------------------


Installation Successful.

Next Steps:

1. Access the About Application Server PE 9 welcome page at:
file:///opt/java/ee5/sdk/docs/about.html

2. Start the Application Server by executing:
/opt/java/ee5/sdk/bin/asadmin start-domain domain1

3. Start the Admin Console:
http://localhost:4848

Java 6 on Fedora, CentOS5, RHEL5 needs compat-libstdc++-33.i386

I downloaded Java 1.6 with EE 5 and tried to install it on CentOS 5. But it complained about libstdc++.

./java_ee_sdk-5_01-linux.bin: error while loading shared libraries: libstdc++.so.5: cannot open shared object file: No such file or directory

I have libstdc++ installed, but have /usr/lib/libstdc++.so.6.0.8

Sun has a bug on it.

Aparently, Fedora, Redhat, and CentOS are not supported. What they needed was to statically link. The solution is to install a previous version of libstdc++ (libstdc++.s0.5). You can do that with yum:

yum install compat-libstdc++-33.i386

and then run the java installer:

./java_ee_sdk-5_01-linux.bin

--

I got Java 1.6 installed, but java_ee_sdk-5 is still hanging:

see http://docs.sun.com/app/docs/doc/820-0135/gazry?a=view

export JAVA_HOME=/usr/java/jdk1.6.0_03

Bugzilla, CPAN, CentOS, VMWare, VPSLAND

I've been fighting with VMWare, CPAN and CentOS this week.

Okay, really only fighting with VPSLAND, who say they've fixed my account and allow me to re-order. Wrongo. Account not fixed. Ability to order (even with a new account) not possible. So I guess I have to look for another hosting provider.


What I've been doing in the mean time is setting up a VMWare server instance of CentoOS 5. My first goal is getting a development version of Bugzilla running so I can start on writing some tests and dig into improving the Bugzilla Webservice module. That was three days ago.

To be fair, I've been sick, and yesterday I mostly played Tower Defense, a cool flash strategy game. Possibly my new favorite after Dicewars.

VMWare has been flaky with network connections, making it difficult to download dependencies. Part of the problem is my piece of junk Gateway MT3423 laptop, combined with Windows Vista Ultimate, which has made losing wifi a frequent occurence. The other part is with either VMWare or Vista, or the two in cohoots, dropping the connection. It seems VMWare is unable to establish a connection after Windows loses it and regains it. Not without a hard boot at least. I plugged into the router today to fix that, and have been merrily downloading, building, and testing all day.

That's right. It took all day to resolve the dependencies, using CPAN, Yum, RPMFind, and building from scratch when necessary. This is not a process I look forward to with client sites. I still don't have mod_php or java set up.

I wanted to have hand compiled perl,php,apache, apache modules, perl modules, etc. that I could copy from my CentOS VM to a production site, but that's not happening yet. I'm itching to do some development.

CPAN truly is a fragile monstrosity that makes DLL hell look like Nirvana and RPM hell like Moksha. It's the number one reason not to use Perl. But unfortunately, also the number noe reason to use Perl.

I have a working Bugzilla on Windows, using PPMs, but 001compile.t fails on install_module.pl and CPAN.pm. I talked to someone on #mozwebtools about it, but didn't get any info. I wouldn't want to support a Windows/ActivePerl install either. It looks to me like it's just coded to use 'cpan install', and not 'ppm install.' I could have got by, but want to have as clean a test environment as possible for when I ask questions.

Once I got all the bugzilla perl dependencies installed (XML::Parser has circular dependencies that can't be resolved vi CPAN, Chart::Base apparently wasn't on any of the mirrors, and GD just plain doesn't work), then I ran into what I thought was a breaker. It looked like VMWare was up to no good with the network interface, but I guessed, and discovered that CentOS has restrictive firewall rules that prevent anything but localhost from accessing all but port 22 (SSH) and ping requests. I just turned IPTables off for now, and then, after some messing with httpd.conf, I finally got bugzilla working.

Email is next. I can send email from the command line now, but bugzilla didn't send an admin email.

Saturday, December 29, 2007

VPSLAND trouble

I purchased a new VPS for qa-site.com through VPSLAND. I've been happy with them in the past. I'd like to continue with them. But they don't seem to want my business.

I signed up on Thursday, and got 5 email notices (twice) basically saying they'd gotten my order and processed my payment.

There was some trouble initially with signing up, because it wouldn't let me sign up with my existing account. That may be my fault. I may not have had the right username or password. Okay, in the spirit of getting things done, I signed up for a new account, using my business credit card. And got the aformentioned notices that everything was fine.

But I never got any notice of being provisioned. I never got an IP address and password to log in. I wrote VPSLAND a kind note asking how long it takes to provision, and by the way, I have another account with them, can I combine them?

I got back several evasive responses in which they finally said my order had been deleted as suspected fraud. Now realize, I would have never even learned of this if I hadn't prodded them. So apparently their policy was to silently discard orders that they think might be fraud. To be fair, they claimed in the support email that the card had been refunded, but though I received 5 notifications (twice) that it had been billed, I have yet to receive any official notification that it's been refunded. I don't see anything on my card statement yet.

Not that I want it refunded. I'd much prefer the service, as I've been quite happy with VPSLAND in the past. And I'd really like to get started on QA-SITE. As of last night I'd corresponded with VPSLAND support several times attempting to figure out just what happened and what I can do to help correct the problem. Finally I received notice that it was "escalated."

I don't know what could have triggered a "suspected fraud" other than if they logged my IP Address it would show that I am nowhere near the zip code where my credit card bill is sent. It goes to Bellevue, Washington and I'm living in Ecuador. Okay, whatever. They have a right to determine what they believe might be fraud, and decide what risks they are willing to take. I'd rather there were more vendors more cautious out there on the internet, and with pressure, maybe the credit card companies could come up with a more reliable system.

But the lack of professionalism in notifying me, the ambiguity of the responses, and the "escalation" to what appears to be a black hole (to be fair, it's been less than 24 hours) are disheartening.

Apparently, the real problem is that VPSLAND has "upgraded" their billing system, and possibly my existing account has been lost in the transition. Fine, we all make mistakes. But I just want my new account working. We can sort out billing irregularities, consolidate my accounts, etc. later.

Tell me why I was suspected of fraud and what I can do to remedy it. If it's not possible, fine, I'll take my business elsewhere, but I'd really rather stay.

Moving on to the next phase -- QA Site

This past week I've been procrastinating. I have the One-Shore.com website (sort of) complete. It's good enough for now. It's untested on older browsers (including IE 6) and it's not the prettiest thing, and the copy is terrible. But it's there. I have what I want, it just needs edited.

Now the thing to do is get a working QA Site (a demo at least.) I've been putting it off. To be fair, Christmas has intervened. I decided about a week ago that I should do it like a real project - from concept, to proposal, to requirements, schedule, budget, on down to deployment. At that point I got lost in the process. After all, part of what I want to do is identify (or build) tools to help me do that sort of thing.

Speaking of which, something else I could be doing when I'm procrastinating is writing blog posts and wiki entries on tools and techniques.

I'm suffering from paralysis on what frameworks and tools to use as well. Should I use a project management suite like dotProject or keep it lightweight? Should I use bugzilla or trac? Should I use django, rails, php (which php framework), or java (which java framework)? Should I write my own framework (in which language?) Which tools should I promote? Should I write my own test case tool? How much should I do up front -- dashboard, web services, virtual appliance?
Should I be spending time looking for customers? Where?

Stuff like that. I need discipline. I need scoping. So that's what I'm going to do. Treat it like I would any other project. Only I'm the CEO, graphic designer, copy writer, developer, tester, project manager, sysadmin, accountant, and consultant.

I'll start with ONE RFP. Written by myself. I started to say "a series of RPFs" but then changed that. Now I'm changing it again. I'll write two. One as a potential perfect customer (knowing what I want to build.) And another as an internal RFP to implement that proposed solution. A "concept" document will proceed the internal RFP.

And then I'll do scoping. I'll come up with a list of features, and then pick the ones I can implement at first. That will (hopefully) be based on the fictional customer's RFP, but with an eye towards expansion to meet future functionality.

The next step is a project plan. How am I going to implement it? What tools am I going to use? It's okay if I don't like what I choose. After a month (or three), I can change. I'll give myself a schedule, and a budget.

After that, then is the design. I'll decide architecture and requirements. I'll document them according to my project plan. If my plan says word docs on a local file, fine. If it says dotProject, or a wiki, fine. I can change it. But try one and stick to it for a month. If I decide django's for sucks, or I really want to build my own PHP framework, fine -- but stick with something. Get a 1.0 out there. Make it a really small 1.0 There's always 2.0.

Plan monthly or quarterly. I think eventually quarterly will be a better fit. But when I'm by myself, monthly is fine. There will be a lot of rapid changes in tools, process, design, features. Release weekly or more. Document it, blog it.

There. Now I'm off to write my concept and fictional RFP. And read a bit more about writing proposals. That's what I'm doing today.

Tuesday, December 18, 2007

Blogs, Forum, Wiki and nav links complete and RESTful on oneshore.com

The One Shore blogs are set up and working. With the correct URLs. As are the forums and wiki. And the about, services, and contact pages. All the menu links work (except demo) as well as a few extra. They are RESTful and pretty (mostly) and are bookmarkable and seachable. They have content. The wiki still has a lot of subjects to fill in, however. http://one-shore.com and http://www.one-shore.com behave the same. They are all under the same document root (through aliases) so migrating from dev->test->prod shouldn't be as much work. VirtualHosts for wiki, forum, blogs, and demo exist but have been deprecated. I'm not sure I like that, but based on Kelsey's opinion I put them under. It does make navigation cleaner, because you can go from http://www.one-shore.com/blogs/company to http://www.one-shore.com/about/one-shore by just editing the path, without messing with the hostname. This loses the independent logging and (easier) access restrictions (and ssl) that virtual hosts have. We'll see.

So here's a quick site map:

http://one-shore.com == http://www.one-shore.com
http://one-shore.com/about
http://one-shore.com/about/one-shore
http://one-shore.com/about/aaron-evans
http://one-shore.com/blogs/company
http://one-shore.com/blogs/technology
http://one-shore.com/blogs/news
http://one-shore.com/blogs/work
http://one-shore.com/wiki
http://one-shore.com/wiki/tools
http://one-shore.com/wiki/qa
http://one-shore.com/wiki/development
http://one-shore.com/wiki/pm
http://one-shore.com/wiki/biz
http://one-shore.com/forum
http://one-shore.com/forum/f2-Shore.html
http://one-shore.com/forum/f15-Main-Support-Forum.html
http://one-shore.com/forum/f18-Tools.html
http://one-shore.com/forum/f20-Techniques.html
http://one-shore.com/demos
http://demo.one-shore.com/bugzilla
http://demo.one-shore.com/mantis
http://demo.one-shore.com/trac
http://demo.one-shore.com/qatraq
http://demo.one-shore.com/testopia
http://demo.one-shore.com/docuwiki
http://demo.one-shore.com/knowledgetree
http://one-shore.com/services
http://one-shore.com/services/qa-site
http://one-shore.com/services/consulting
http://one-shore.com/services/staffing
http://one-shore.com/services/web_development
http://one-shore.com/contact
http://one-shore.com/contact/sales
http://one-shore.com/contact/sales/qa-site
http://one-shore.com/contact/support
http://one-shore.com/contact/careers
http://one-shore.com/contact/investors
http://one-shore.com/contact/partners

Also the following external links:
http://demo.qa-site.com/
http://fijiaaron.wordpress.com
http://fijiaaron.blogspot.com
http://boneshore.runboard.com/ == http://www.runboard.com/boneshore
http://bqasite.runboard.com/ == http://www.runboard.com/bqasite

Plus lots of tools in the wiki, and more forum pages. As you can see, the forum URLs aren't the best. Getting it merely took a tedious series of cut and paste steps from a tutorial on adminpicks.com Let me just say, punBB is spaghetti and I salute the brave soul who created the patch. Actually, a patch wouldn't have worked because there's been some change since then.

So now my site is structually complete, and restful, it's time for me to get some rest, and then get down to the real work. Plus, there's going to need to be some backporting because I did a lot of stuff on the fly in production, and I don't have it mirrored in dev. Like the nav links, I think. And any blog, wiki, or forum content. And I still need to test the contact forms.

Wednesday, December 12, 2007

Site and Business progressing

Things are starting to come together.

The one-shore.com website is starting to get fleshed out. Yesterday I got the wiki (dokuwiki) installed and created all the tools links. Today I got the forum (punBB) installed and created categories.

The wiki is now more up to date than the tools links. I still need to enable rewrite on the wiki and give it one shore branding (colors, logo, and nav links to return to the rest of the site.) I also want to add the tools links to the side for easy navigation, as well as a link cloud, by most visited, and post info on the home page such as most popular, and latest updates.

The forum is fully branded with a bevy of sample categories and forums. I think I need to notch it down and just make a few major categories. One Shore, Tools, QA Site, Support, perhaps combining the last two. It goes Category->Forum->Topic->Post --- which is one more level than I anticipated creating it. Also, it seems that all the way down to topics, only the administrator can post. I'd like registered members to create topics, for things like adding a tool or posting a support question.

Multiple blogs are available on dev, but unbranded as yet. I may continue to use my wordpress.com and blogspot.com blogs for now, because it provides an offsite location and the possibility that someone may stumble upon it. Eventually I may mirror locally or migrate.

I also created forums on runboard.com, which offers an off-site location, but may be wasted space. It will be a good back up if I have to bring my own forums down.

Yesterday I setup a paypal account so I can receive payments through that. Today I got a skypein number: 206-801-1701, so I can receive calls from the US easier. Only sound (and hence skype) doesn't work on my computer, at least since "upgrading" to Vista Ultimate (it swears there are no sound devices) from Vista Home --so I will have to use Kelsey's for it. I think I may want to go with TalkPlus eventually.

Skypein is $18 for 3 months. It's a good number (though the geek in me wanted 1734) and I'd hate to change it if I start to depend on it. Talkplus is $6.99/month plus 11 cents a minute and can ring my land or cell number here in Ecuador.

The deciding reason I got a number is because I sent an email to be listed on bugzilla's paid support. I also listed as a user. And as a user of Mantis (I use mantis for the one-shore site development and Bugzilla for the qa-site dashboard and integration.) I also have demos of both, and will support either for customer QA sites. I also made a note on Dokuwiki about using them, and emailed Knowledgetree about using their product. Knowledgetree told me about sourceforge, and I registered on there, though I need to look and see what tools I support are on Sourceforge and add that to my sourceforge market profile. I don't think bugzilla or dokuwiki are.

Tomorrow I hope to fix all the links and get the about, contact, and services pages finally up.

I will be visiting the computer school at San Blas square hopefully this week and eventually the technical college looking for people with good english and computer skills looking to join one-shore. I'll also try to network with people online and contact open source project leaders, and maybe company like Knowledgetree, but also, for instance, Atlassian. Crowd is staring to look very appealing to me, though I want to have the option of java-free VPS accounts for lightweight users.

I'l post on craigslist, guru, and look for other places soon. I'm probably not willing to cold call or spam businesses, but need to find out how to talk to someone who might really be interested.

Thursday, November 22, 2007

Code for corners

By the way, here's the code for the rounded corners I'm using:

.outer {display:block;}
.outer *{
display:block;
height:1px;
overflow:hidden;
font-size:.01em;
background:#DDDDDD;
}
.outer1 {
margin-left: 3px;
margin-right: 3px;
padding-right: 1px;
padding-left: 1px;
border-left: 1px solid #686868;
border-right: 1px solid #686868;
background:#aaaaaa;
}
.outer2 {
margin-left: 1px;
margin-right: 1px;
padding-right: 1px;
padding-left: 1px;
border-left: 1px solid #252525;
border-right: 1px solid #252525;
background:#b6b6b6;
}
.outer3 {
margin-left:1px;
margin-right:1px;
border-left:1px solid #b6b6b6;
border-right:1px solid #b6b6b6;;
}
.outer4 {
border-left: 1px solid #686868;
border-right: 1px solid #686868;
}
.outer5 {
border-left: 1px solid #aaaaaa;
border-right: 1px solid #aaaaaa;
}
.outerfg {background-color: #DDDDDD;}



.inner {display:block}
.inner *{
display:block;
height:1px;
overflow:hidden;
font-size:.01em;
background:#111111;
}
.inner1 {
margin-left:3px;
margin-right:3px;
padding-left:1px;
padding-right:1px;
border-left:1px solid #8e8e8e;
border-right:1px solid #8e8e8e;
background:#484848;
}
.inner2 {
margin-left:1px;
margin-right:1px;
padding-right:1px;
padding-left:1px;
border-left:1px solid #d7d7d7;
border-right:1px solid #d7d7d7;
background:#3a3a3a;
}
.inner3{
margin-left:1px;
margin-right:1px;
border-left:1px solid #3a3a3a;
border-right:1px solid #3a3a3a;
}
.inner4{
border-left:1px solid #8e8e8e;
border-right:1px solid #8e8e8e;
}
.inner5{
border-left:1px solid #484848;
border-right:1px solid #484848;
}
.innerfg{background:#111111;}



Thanks alot to the guys who created Nifty and Spiffy corners. Alessandro Fulciniti & Greg Johnson and everyone else who shares code and ideas. That's what it's all about.



Site design

I'm working on a site for one-shore, and have wasted a few days on design, but I don't know if I like it. I started with the idea of using a picture of a "shore." Here is my concept drawing:



I wasted a lot of time looking through photos on stock photo sights, flickr, and google images, but couldn't find anything quite right. I looked through a bunch of my travel photos and settled on one -- a sunset somewhere in Fiji:



It's not perfect but it'll have to do for now. Alternately I'd like a wave rolling from right to left or a smooth rock beach. Maybe have a series of images rotate like banner ads.

Getting rounded corners was a nice web 2.0 touch, and I found a cool trick with CSS on SpiffyCorners.



I decided a black background works better for the image:



Putting one rounded corner div inside another gave me the outline effect here:



I discovered with relative position -2 I could get it thinner on one side, but then it'd be thicker on the other:



Finally I figured out how to get it only 3 pixels wide, by interleaving the first 2 elements of the inner rounded box with the latter:



That took a full day. Now I'm working on the layout of the rest of the page. I was so proud of my rounded corners that I wanted to use them everywhere:



But now I'm thinking less is more. Too many corners is too "cute" and I want to look more "serious."




I'm still not sure about the black background, especially since the wiki and other tools will be a major part of the site, and it will be jarring to change. Next I'll try to turn this into a wordpress theme.