peace on the web, too

My first impressions after a couple of days of testing 2 mail subscriptions services. I did a comparison between R|mail service and RssFwd one.

R|mail post email

From: Rmail
To: …
Date: Sep 15, 2005 1:24 PM
Subject: Let’s … share! - mapopescu was here, too …

About the Web 2.0 need for API discovery, sharing and integration.

web dev tagging
Click here to view the original Webpage.

---

Sent using R|mail.
Click here to unsubscribe.

RssFwd.com post email

From: mapopescu was here, too …
To: …
Date: Sep 15, 2005 1:20 PM
Subject: Let’s … share!

Recently we saw a couple of blogs/people trying to put up a Web 2.0 APIs map. Just take a look at ProgrammableWeb.com and Web Service Finder (via ReadWriteWeb) and see for yourself..

[whole article content followed]

We are looking on something, but we don’t know what is it just yet.

Marius Popescu Thu, 15 Sep 2005 11:03:56 +0000

---

Source: http://mapopescu.blogsome.com/2005/09/15/api-sharing/
To unsubscribe from this feed, click here
To manage other subscriptions, click here

Powered by RssFwd, sponsored by Sproutit.com: software for small business. Find out more

My first impressions

R|mail pros:

• very concise
• showing the summary (have you an option for sending full text?)
• showing article’s tags (excellent). I don’t know if I like that they point at your site, though!

R|mail cons:

• the link to the article should be at the beginning, not the end of the email, especially because you are sending only the summary.
• I like better the RssFwd’s “From” content
• from the subject content you don’t know that it is a blog sending you an email, the name of the blog is rather lost at the end
• your website is kind of complicated for me (and I am a web power user), perhaps a usability test won’t harm
• your “unsubscribe” feature seems kind of too brutal, a confirmation would be necessary for me
• I don’t like the proposal for integration (the design)

RssFwd pros:

• “From” is better
• the article link is more visible, even if it’s at the end (as the email contains the whole article, the link becomes less important too).
• you can see all your subscriptions
• when subscribing you can see the feed’s contents before subscribing, on their page
• their site is more simple (and usable)
• you can subscribe using only a link

RssFwd cons:

• the site is slower
• no feedburner.com statistics

Conclusion

Sorry if I seem more attracted to RssFwd. For now this is the truth.

PS A special note: R|mail appears on my feed provider (feedburner.com) while RssFwd doesn’t.
I don’t know yet why, perhaps is their way of doing things. This could become important.

PS2 R|mail is a service provided by a single person. RssFwd started the same, now it seems to be sponsored by Sproutit.com. RssFwd is open source too.

PS3 I evaluated also FeedBlitz.com, but it look too orange to me, perhaps a much better service like the two presented here, as it seems to be a company, but way to complicated for me :)

In the early days (Delphi 1 to 4) I didn’t had a great interest in the Delphi Internet development proposals. I think it was Delphi 4 that proposed for the first time *Internet Express*. Delphi 5 introduced *WebBroker*. Delphi 6 added the *Websnap* layer to WebBroker. And then … silence. Oh, in the most recent products (Delphi 8 and 2005) there is ASP.NET (with code behind in Delphi), of course.

It is good to mention the two most notable third-party web frameworks (built on top of WebBroker and/or Websnap): “ExpressWebFramework”:http://www.devexpress.com/Products/VCL/EWF/ and “IntraWeb”:http://www.atozed.com/intraWeb/.

Personally, I have used Internet Express, WebSnap and ASP.NET. I only took a look at the third party ones.

So, let’s start with *Internet Express*. The central point of the development were the _page producers_. These were components that could take an html file or stream, pass it through a tag replacement filter and feed it to the response. Their design was so simple and generic that they can be used even in desktop applications (creating html dynamic pages as replacements of the classic grids or help pages).

The big thing, looking back at it, it is the InternetExpress’s MIDAS link. In fact, with no much of a hassle, one could write a desktop application and a web application using the same code on the server. By same code understanding the code providing client-side datasets with data from the database. I will get back on this one later, coupled with AJAX (note to self).

*WebBroker* went a lot further and proposed _web modules_ with _actions_. The actions correspond to the URLs that your application will respond to. The big news was that one could compile a CGI, ISAPI or Apache module (dso) with almost the same code, using only different project files. This was a big advance, and a hell of a reason to use Delphi. At the time WebBroker appeared, Borland provided Kylix, a Delphi/C++Builder sibling working on Linux. Add the triple offer (CGI/Isapi/Apache mdules) to Linux compilation and we have something very interesting! This was the time Delphi was in the frontline.

And here comes *Websnap*. Announced in a big way (_the RAD web framework_), Websnap is now a big disappointment. It was built on top of WebBroker and offered a lot of advantages, not loosing anything of the WebBroker’s offer. I must mention that I am still using it until today in my work.
And it was not bad. For starters, you got multiple WebModules, user management base components, session management (in-process by default, easily to expand to database hosted), adapters (layer between datasets and the produced page) etc.
But, man, it was complicated! Anything in it was written using interfaces. Which is not bad. But is very difficult to debug. Try only to debug a CGI. Now imagine ISAPI. That’s cold outside. Get in!

For me, it looks like at that point (between Delphi 6 and 7) it just stopped. At the same time, Delphi guys (read managers) saw .Net and we were left with our hammers and stones… I mean, Kylix stopped at the same time, Websnap too, this is a big chunk of Delphi! The Linux part and the Web part.

Now, we got ASP.NET. It is true, Websnap and WebBroker and Internet Express are still there, in the Win32 version of Delphi 2005. It has quirks, same bugs, same difficulties, no third party support (except those two frameworks mentioned before) - which shows how difficult is to grow something on those components.

Me, I am thinking that a combination of Websnap with InternetExpress and AJAX would (have) kill just any concurrence. They got it all, they just let it slip.

Now back to my Web 2.0 blogs, where nobody mentions Delphi no more ….

Update: some typos, some tenses …

Update 2: As usual, Marcu Cantù gets it sooner than others ;)

Recently we saw a couple of blogs/people trying to put up Web 2.0 API maps. Just take a look at ProgrammableWeb.com and Web Service Finder (via ReadWriteWeb) and see for yourself.

This seems to go in the direction of the Web 1.0 open directories like dmoz.org or google and yahoo own directories. And this is fine and usefull etc. For now at least.

However, from a Web 2.0 point of view, I think we need more. Those API directories need to be alive. We need RSS feeds on them, that’s OK too for starters. But how about discovery, sharing and integration?

I would say those API directories need to have their own APIs! And a common standards would help, I don’t know. But the most important is to have one. And if nothing else, the discovery, share and integration parts should be the most important. Andd let’s not forget to recommend inter-API integration facilities. Like using the ProgrammableWeb.com’s API to search for other API directories. This should be easy, if the other API directories would have one APi for themselved too.

And for us, the mortals, this would make all much less complicated. Need a photo album API? Just select one directory, check their API and let the discovery function gather for you anything needed.

Of course, you would say, which one of the found APIs should I use? Well, my friend, I knwo this will showk you, but you must have your own API that tells you what your social network recommends you. You have one, right? And your fellows networkers have one too, right? And they tag the APIs they use, right? And your network has an API too, which you can use to extract a top of a specific domain APIs, right? …

We are looking on something, but we don’t know what is it just yet.

I am a long time Delphi user. My interest in web development forced me to check other languages or areas that a standard Delphi (or desktop applications) programmer avoids or simply ignores. I am sorry, but this is the reality. Check the Borland Delphi newsgroups and see the amount of discussions on internet-related technologies for yourself. Those newsgroups are simply dying. The ASP.NET ones are still moving because they are new.

I even asked a couple of times if there is anybody using AJAX in Delphi. I had only some responses asking “what is this AJAX thing?” and it just stopped afterwards.

As always when it comes to Delphi web development, I see a hope in Marco Cantù’s books and articles. And now he’s even presenting an AJAX in Delphi session at the 2005 European Borland DevCon, which I will be trying to attend.

Note to self: an article on Delphi Internet technologies history would be usefull.

As a romanian expat I already use all the modern way of comunication when talking to my friends or relatives back in Romania or from abroad. When IM begun, I used Yahoo messenger and MSN from the beginnings. When Skype arrived I used it very soon too. And so was my wife. And so is her family back in Romania. Etc.

What I mean here (and tiying up these readings) is that developing countries (you can say poorer too) like Romania is too are full of early adopters of modern comunication tools. They all use y!m, msn, skype etc. From the first day. Recently (at least in the most developed countries of those cited) there is an explosion of debit and credit cards. I guess this comes with the development. But almost anybody uses them on the web. There is fear but also statistics that block people doing e-commerce.

But what if some already using skype had access to ebay? If she uses SkypeIn too, there is a way of certifying her identity. So ebay customers could trust more a person if they could check its existence (call her, check her credit card existence etc.). So the long tail of e-commerce users (romanians, indians, russians etc) could come to light. They could use e-commerce because the combination of a market (ebay) and personal comunication (skype voice and SkypeIn) could be trustfull.

There is a lot of buzz on the web right now about rss, xml, subscriptions etc. - all forms of subscribing to the content of a site. It all comes down to the ability of a reader (a human one) to subscribe to a content publisher or provider.

Of course, this idea coming from the blogs world or having the more impact on blog readers, it is right now covered with a lot of techie talk (like those 3 letter acronyms on the upper paragraph). I really hate it, and I have a lot of difficulties to explain this - rather simple - subscription idea to a non-tech person.

Why wouldn’t we think more generally? I see this possibilities that are not enough explored for now:

• email subscription:
  We already have email subscription for newsletters and email groups, why not use it? I am testing right now a site providing this service (http://www.kbcafe.com/rmail.aspx) but there are some more offering this. However they seem limited to a newsletter approach. You get emails with articles. This is a passive approach, which can be just fine for starters. But why wouldn’t we have the possibility to use the subscription site as a reader too?
• tag/category/search subscription:
  Almost all blog engines I saw provide automatic feed creation for the whole site. I would need a more dynamic approach: subscribing to a tag/category, to a search on my blog (look a the searched words as to “proposed tags” and you’ll get a new way of discussion), to a “fixed” page (like “about”) etc.
• subscription items look:
  Why on earth the readers available today don’t propose a (default) view of a blog paying respect to the original look (style)? A microformat on such a thing could help perhaps. Anyway, the look of the posts is an issue for, but not limited to, design people mostly.
• subscribe to a static page:
  I don’t know why I can’t subscribe to a “static” site in the first place! The “posts” could contain only the changed content, all the content with the changes highlighted etc.

Almost all the issues I see on this subscription issue come down to this: subscribing to any web content should be simple, non-technical and possible through all the classic or modern ways. Applying this principle would produce new ideas/approaches that would improve further the dialog between the reader and the writer (as my idea of “proposed tags” that I come up with whne writing this post).

I am thinking about or analysing 2 or 3 web applications. Any time I want to hit the road with one or the other there is something that stops me. Sometimes are silly things, human limits etc. Sometimes I think the immensity of the web imposes me. How can you create something that is (you think) unique and usefull?

I was recently reading some articles about the future of programming (The history of physics and the future of computing). I don’t know if I understood everything, but the idea of “bind rather than select” it just fits in when you work on the web.

It makes more sense to just discover things on the web when you are using a service or application, rather than having to specify or select something explicitly. How on earth should I know if I prefer this or that online photo album for storing my pictures? Or why should I “enroll” my site to a search engine list? Aren’t they suppose to search for it?

These all are big questions and I feel I’m missing something when a web service I use doesn’t “discover” stuff for me. I’m not talking about spam, newsletters etc. Think about those long tail right-side offerings you see when you look at a book or CD online review. It’s just normal to get those related artists or writings showe dto you. How could I discover them by myself?

A good idea (that I have on my todo list for every web project) might be to use microformats or something similar. That way, a discovery service could understand why are you linking to something, automatically. But this involves work for the programmer and for the user.

Getting back to the article that started this post, it would be really usefull to have these discovery or binding capabilities right into our web framework, whether is Rails or Delphi or PHP, whatever. Perhaps dynamic languages (like Ruby or python) are more prepared, but the others have the experience on their site. Let’s wait and see …

It is unbelievable. Every single website, mainstream or newcomers, are or will be offering a tagging (labeling, naming, keyword-ing) facility! This means every time I am somewhere on the web I can tag something on someone’s website with some words, saying something I think about a specific item.

This creates something called (tagged;)) folksonomies for the respective websites, dynamic categories, trends in reader’s habits or preferences etc. But how about me? What do I get?

I think all those Web 2.0 (or Web 1.0 going 2.0) companies should think about me and give me something in return of my contribution(s). I think if I am to register some service (like amazon, flickr, whatever) and tag something (which does not necessarily belong to me), I should be able to get a list of my tags and perhaps links to get XMLs or some kind of list containing all the site’s content matching my tags. It seems to me the least decent thing to do. If not, what do I get by tagging something?

This could go as far as getting a list with my search words from my search engines (where I am a registered user, of course). Or links to all the photos matching each of my flickr tags. Or bookmarks I tagged at del.icio.us or furl.net (I know they are already doing it). Etc.

And this could be later integrated into my blog or some kind of private space where I could follow all my tagged content. In a way, if I tag something, I relate to it. It is a 2 way relation. For now, I see mostly the first part: we, the websites, get tags. You, the user, get other’s tags. Not only yours.

PS I guess there are websites already offering this service. I apologize if one cited by me already has it. I am only trying to get real examples.

It would be great if a website could integrate for a logged-in user the pictures found on his hard drive with the ones being uploaded to a server (or more).

The website could present this as an option/feature. The user would have to download a local indexing service for her pictures, this would be the downside of the thing I guess. Perhaps it should use services like Google Desktop … or even Windows indexing engine, I don’t know, but it should be able to be accessed by http on a local port.

Then, when the user logs into the account on the website he can see all uploaded photos (tagges, labelled, chronologically etc.) as well as local photos. Maybe we can push the integration further to link all photos (by dates perhaps) in the same lists, so that it become easy for someone to see that she has some photos not uploaded yet.

Come on, which will be the first Web 2.0 company to offer this? flickr? buzznet? Somebody? I have about 2 Gigs of pictures just dying on my hard drive :p