La potenza della rete sta nella enorme mole di contenuti che produce ogni giorno. Trovare e aggregare notizie è una caratteristica chiave.  Cos’altro è google news se non un aggregatore delle migliori fonti di news disponibili.

Per una organizzazione estesa come la Sapienza, la possibilità di aggregare contenuti, istituzionali e generati dagli utenti è strategica.  Ci sono strumenti gratutiti e potenti a disposizione dei web master per scandagliare, filtrare e aggregare i contenuti in rete in maniera da fornire un servizi di informazione completissimi.

Yahoo Pipes è tra i più strabilianti.  Ripubblico una recensione che ho fatto qualche tempo fa.

Cosa è Yahoo Pipes

Secondo Tim O’Reilly addirittura:

is a milestone in the history of the internet.

Una pietra miliare nella storia di Internet. Forse è troppo ma di sicuro Pipes è un passo avanti notevole nella direzione della semplificazione dei servizi web e della interoperabilità.

E’ un servizio web con un’interfaccia grafica sofisticata che permette, agli utenti senza particolari conoscenze di programmazione, di assemblare web services abbastanza complessi.

Si possono prendere dati da fonti aperte, pubblicate attraverso i formati di RSS (Really Simple Syndication) o accessibili attraverso API (Application Programming Interfaces) , manipolarle ed aggregarle usando operatori logici, funzioni per filtrare i contenuti, concatenare stringhe, estrarre informazioni utili per la georeferenziazione, introdurre informazioni da parte degli utenti.

Il risultato è un servizio web personalizzato, disponibile in formato RSS utile per essere integrato nelle proprie applicazioni web.

Un esempio.

Si possono prendere le notizie pubblicate su un quotidiano on line, filtrarne i contenuti ed associale con le fotografie che su flickr sono state classificate con le stesse parole chiave fornite da una funzione di Pipes per analisi testuale che le estrae dai contenuti selezionati.

Questo esempio prende alcune parole chiave delle notizie del New York Times e le associa alle fotografie di Flickr.

http://pipes.yahoo.com/pipes/pipe.info?_id=vvW1cD212xGMiR9aqu5lkA

tutto in maniera automatica.

Un’altra applicazione per cui Yahoo Pipes è particolarmente utile è il news mastering. Selezionare diverse fonti filtrarle e fonderle in un unico feed è semplice e permette di immetere nelle proprie applicazioni on line notizie di qualità mantenedo un certo grado di controllo e di coerenza con la propria linea editoriale. Pipes può essere usato come utlile strumento in combinazione con altri servizi per il news mastering come MySyndicaat. Con MySyndicaat per creare il codice html o php da inserire nel proprio sito per mostrare la propria selezione di notizie. Il rislutato si sarebbe potuto raggiungere anche con il solo MySyndacaat ma per ogni parola chiave è necesario inserie i feed di tutte le fonti utlizzate. Con Yahoo Pipes invece queste sono selezionate solo la prima volta e poi basta inserire la parola chiave per avere l’output.
In generale Yahoo Pipes è un utile struemnto per assemblare gli output di tutti i servizi Web che vengono distribuiti in formato RSS feed o con API.
Il limite è la fantasia

La comunità di sviluppo si trasforma in social network

Pipe unisce un sofisticato ambiente grafico solitamente dedicato alla programmazione (assomiglia alle interfacce visuali per la programmazione ad oggetti e ai database relazionali) e lo consegna ad una rete di utenti più estesa. La comunità di sviluppo di trasforma in social network accogliendo anche quegli utenti che non hanno una conoscenza dei linguaggi di programmazione.

Una volta creato un “Pipe” infatti lo si può pubblicare e gli altri utenti lo possono clonare modificare o utilizzare per costruirci sopra altri servizi. Un pò come usare i lego.

Il concetto di “Pipeline” è derivato dal mondo Unix. Nel sistema operativo più famoso del mondo come nei suoi figli (Linux) è possibile convogliare l’output di diverse applicazioni direttamente in altre applicazioni del sistema, costruendo comandi complessi (chiamati appunto Pipelines) e automatizzando compiti complessi.

La stessa cosa fa Yahoo Pipes utilizzando i web services invece delle applicazioni.
Il social network per lo sviluppo ed il test di servizi web

Semplicità d’uso e condivisione quindi sono i punti di forza per questo servizio che promette di dare una bella accelerazione al già prolifico fiorire di servizi Web 2.0. Dal punto di vista dello sviluppo e del successo del servizio proprio l’approccio “social” è l’aspetto più rilevante di questo esperimento di Yahoo.

Invece di percorrere la strada del modello di sviluppo Open Source con una “centrale” per lo sviluppo e per il rilascio di aggiornamenti, patch e documentazione che solitamente porta ad una serie di progetti clone o poco differenti (forking) seguendo gli umori della comunità di sviluppo, Yahoo ha optato per un approccio “embracing failure”. In parole povere cercando di esternalizzare ridurre i costi del fallimento di una idea (traducendo come meglio possibile, costs of failure).

La forza dei progetti open source e quella maggiore delle idee prodotte e realizzate dai social networks risiede proprio in questa capacità. Le imprese hanno costi strutturati, anche la sola progettazione e scelta di diverse strategie di sviluppo comportano dei costi e dei tempi di approccio al mercato. I risultati di servizi come in presenza di servizi flessibili e con un livello di astrazione elevato come Yahoo Pipes i costi di sperimentazioni di tutte le possibili combinazioni che portano ad un prodotto o servizio di successo sono elevati.

Lasciare alla rete la possibilità di combinare e ricombinare gli elementi che lo costituiscono è senza dubbio la migliore strategia. Per gli individui che partecipano volontariamente al social network questi costi sono molto più ridotti per non dire nulli: il loro obiettivo non è il successo del prodotto ma la sperimentazione, l’apprendimento ed eventualmente lo sfruttamento di una idea vincente o un servizio attraente per il mercato.

Diciamo che in questo senso lo scambio di mercato in questione è infrastruttura e tecnologia a disposizione dal lato di chi offre la piattaforma, lavoro, creatività ed esperienza dal lato di chi partecipa al progetto. Un terreno nuovo di scambio di mercato il cui valore è difficile da misurare usando esclusivamente il parametro monetario.

Uno schema comunque funzionante ed in certi casi dalla forza prodigiosa e capace di creare innovazione.

La differenza sta nel numero di persone che Yahoo Pipes rischia di riuscire a coinvolgere in una attività solitamente ristretta ai cosiddetti geek, un numero più esteso di persone. L’esperienza in programmazione richiesta è limitata quello che serve è fantasia e logica (la famosa business logic delle architetture 3 tier).

Dietro l’interfaccia:

Dal punto di vista tecnico il servizio non è altrettanto innovativo. Strumenti per il mash up di dati ce ne sono diversi ma richiedono competenze da programmatore. Vediamone alcuni:

• Plagger: un “dirottatore” di Rss feed opensource creato da Tatsuhiko Miyagawa che permette di fare più o meno tutte le stese cose che fa pipes ma senza avere l’interfaccia grafica.
• My Syndicaat: il miglior sistema per aggregare diversi feed e creare un servizio personalizzato Si possono filtrare i feed, eliminare i duplicati, scegliere un lay-out per l’integrazione e diverse tecnologie di scripting;
• Dapper: The Data Mapper, è un servizio che permette di raccogliere (grab) dati da tutti i siti, strutturali secondo un proprio modello e renderli disponibili come feed RSS. Meno sofisticato di Yahoo Pipes ha però il vantaggio di funzionare anche con i siti che non mettono a disposizione API o feed. Seguendo il DOM (document Object Model) raccoglie i dati da qualsiasi risorsa web. Certo il risultato non è dinamico ma comunque utile. Costruire directory e servizi con frequenza di aggiornamento abbastanza lunga diventa facilissimo. Può essere utilizzato come integrazione di Yahoo Pipes che non permette fare il Grab di siti che non offrono formati RSS o API.
• OpenKapow: E’ praticamente Yahoo Pipes ma necessita di un client da installare. Non l’ho provato direttamente ma l’interfaccia grafica è meno accattivante di quella di Pipes

Conclusioni:

Tutti questi servizi che di fatto permettono la replica di servizi già esistenti e la loro combinazione di fatto favorendo la ridondanza portano a fare diverse considerazioni:

Per tutti i servizi che fungono da fonte di informazione costituiscono un sistema utile per aumentare la loro posizione nei motori di ricerca. La link popularity cresce a dismisura (con l’effetto collaterale che se il pagerank del sito che li utilizza è basso c’è il rischio che risulti penalizzante. Generalmente si rivolgono comunque ad un target internet friendly con la probabilità di finire su siti con un page rank intorno al 5;molto elevata).
Se è vero che “Content is The King”, trovare contenuti di valore e utilizzarli è sempre più facile. La variabile critiche per il successo di un sito su internet diventano quindi altre: la search engine optimisation, la user experience, il coinvolgimento degli utenti nella produzione di contenuti. Per chi progetta applicazioni e servizi internet quindi l’imperativo è “creare un ambiente “confortevole” per gli utenti e per i motori di ricerca. Anche se i contenuti sono disponibili altrove gli utenti torneranno dove si trovano a loro agio. La conseguenza sulla tutela dei propri contenuti è rilevante.
Cercare di difendere la proprietà ed i diritti sui contenuti diventa troppo costoso rispetto all’esigenza opportunità di scambiarlo e arricchire la propria applicazione o il proprio servizio. La condivisione porta maggiori vantaggi e contribuisce a diffondere il marchio e a far conoscere la propria attività.
L’ultima considerazione la lascio per il nostro paese. Come al solito in ritardo nel comprendere l’evoluzione rapidissima dell’economia della rete ma con delle opportunità. Con fatica ma grazie ad un intervento pubblico esistono numerose fonti di informazione istituzionale con contenuti che se ben assemblati potrebbero generare valore. Dati che in qualche maniera sono “pubblici” proprio perché offerti da siti della pubblica amministrazione. Un patrimonio che resta però confinato nei domini .it, e .gov. Sarebbe ora di aprire le porte, distribuire i dati attraverso RSS o meglio fornendo API ai diversi servizi. L’Internet Italiana ne beneficerebbe e subirebbe una salutare accelerazione che creerebbe valore.

fonti:

• Anil Danish blog
• Harvard business On Line
• Tim O’ Reilly Radar

Sometimes you need to have a quick solution for a rather complicated procedure. I, for example, needed to submit my custom YouTube playlist to Blinx for indexing. Blinx requires users to utilize RSS 2.0 with MediaRSS AND enclosure tags. By default, YouTube’s RSS feed does not include enclosure tags. What to do?

Well, you could learn YouTube API and create your own version of the feed. Or you could use youTube playlist yahoo pipe and thank it’s creator for solving the problem before it even started. All you need is your YouTube playlist ID and voila! Yahoo Pipes generates the augmented feed, you submit that feed to Blinx, and you’re done!

Sidewiki announced some changes to the API a couple of days ago.

The most important one being the parameter:
includeLessUseful=true

This gets around the idea that someone at Google determines what should be included in a feed and what isn’t.

Another fix is that the API can now picks up an entries across a domain. I built a quick Yahoo pipe that demonstrates this although I also see that sidewikirss has already been updated.

What I’m still waiting for Google to provide is some form of search across all Sidewiki entries regardless of which domain they are on.

In a previous post I spoke about RSS feeds and the way reader websites or software allow users to draw together their desired content so it can all be viewed in one place. I mentioned Google Reader but I have since come across an even better website that can do much more than simply consolidate content. It’s called Yahoo Pipes and can be found here.

Yahoo Pipes is a free, web-based service that allows users to ‘remix’ various website feeds by applying filters, doing string searches, obtaining inputs and sending them on to be processed by other modules, manipulating URLs and much more without the users necessarily having to know how to program or any lines of code. Instead users can build these feed processors, or ‘pipes’, by dragging and dropping elements on a canvas, giving them different attributes and connecting them together. It might sound complicated but have a look at this introductory video and you’ll see the basic process is actually quite simple. The completed pipes can be shared with others and added to a user’s Yahoo page. These pipes are often referred to as ‘mashups‘, which is a term generally used to describe applications that mix and match data and services from two or more sources to create a new dynamic entity or service.

As an example, I created a pipe to filter the feed from this blog and omit all posts that contain the word ‘task’. To view the results, go to http://pipes.yahoo.com/pipes/pipe.info?_id=adea8477cbf0c8654dd3e0f87c07a248

This wiki has been launched over a year ago, but the approach to data piping is still relevant today. Go, go Yahoo Pipes!

Start Data Plumbing With Yahoo Pipes – Webmonkey

As APIs and RSS feeds become more commonplace, we look for ways to mash them together. Even if you are a programmer, it can be a tedious process to write scripts that download and parse XML files.

Yahoo Pipes solves the problem by providing an easy, graphical way to interact with data. No command line or shell scripts necessary!

Having written about searching Sidewiki a couple of weeks ago, I’ve been  reading a bit more about it and I’ve also been been playing about with the API.

Google announced some changes to Sidewiki in mid-November, I’ve therefore marked up deletions to this post of things that have changed.

I’m approaching Sidewiki from the perspective of someone who is involved with a brand and wants to ensure I understand Sidewiki’s potential as another conversation channel before/in case it is adopted in a wide-scale way.

In order to help me do this  I created a couple of Yahoo Pipes that use the Sidewiki API – one takes a URL and checks the API for Sidewiki entries (in a similar way to Claude Vedovini’s  Sidewikirss) and then another which takes a list of URL’s in a CSV file.  I then thought about another pipe using an RSS feed (e.g. to easily pick up all comments across a blog) however by that stage I discovered the limitations of the API.  I saw someone selling commercial software that could be useful but they recognised the limitations – they were the people that put me on to the idea that the API only surfaces comments that have been liked.  From my reading around the subject I  think that the API is actually much more severe than this – basing what it shows on an alogorithm.  As far as I can tell this means that the only way of checking your site for comments is by visiting every page (and every variation of each URI – something that is impossible to do – see parameter-ised URI below) with a version of Firefox or IE with the Google ToolBar installed and Sidewiki operating!

My frustrations with the API combined with some of the very negative posts I had read and I started wondering whether in releasing Sidewiki Google had broken their “Do no evil” mantra. This is  something that would sadden me as I’ve always been a bit of a Google fanboy.

I’m not naive enough to think that a brand is in control of how it is perceived, nor that there is anyway that you can follow every conversation involving your brand.  So I thoughtsome more and decided, that excepting that potentially rogue content can be positioned next to your own within the browser, Sidewiki comments are no different from those found on Facebook (a truly walled garden in which conversations can easily be missed).  Then I came across this excellent Neville Hobson post and I felt that much of my thunder had been stolen.  However Neville’s post raises some interesting issues for me:

1. The idea of Sidewiki as a commenting platform.  I don’t think this is something that brand/website owners should encourage mainly because comment monitoring is so difficult.  For example Neville invites people to leave a Sidewiki comment on http://www.nevillehobson.com/.  People might see this and assume that Neville has got the monitoring in place for http://www.nevillehobson.com/2009/10/25/opportunity-knocks-with-google-sidewiki/ or for http://www.nevillehobson.com/2009/10/25/opportunity-knocks-with-google-sidewiki/#comments and other URLs.  However such monitoring may only be visible from your website logs – and then only if you know what you are looking for, e.g I’ve commented on this:. http://malleablemusings.wordpress.com/2009/10/10/searching-sidewiki/?putsomethingherejusttoshowthatit’snotthesamepagewithoutthisparamete.
2. Google have provided a mechanism for page owners to provide a top placed comment in the side wiki.  However you need to do this on every page on your site.  What’s more there is no such option on public sites in which your brands might be investing, e.g. think how much time and energy you might push into your Facebook Fan pages or You Tube Channels etc. (e.g. this video from UNSW has a Sidewiki entry).

I know that there are Sidewiki Policies which supposedly provides some protection against vandalism (I don’t like using that word – as I see them as comments that need to be responded to). However for me the bigger issue is that Sidewiki isn’t searchable (it looks like Google isn’t indexing sidewiki entries and may be removing entries that have already been indexed) and then there are the issues with the API highlighted above.

In it’s current state like others I’m left confused about Sidewiki. At the moment I think movement in this direction is inevitable but can only see it being gamed and as someone who is involved with a brand it scares the hell out of me. I guess that we all just need to keep watching.

Pour mémoire : j’écris désormais “pypes” plutôt que “[yahoo] pipes” pour contraindre à une lecture anglicisante et accentuée du mot, et entraver toute lecture désordonnée du terme. Je ne l’emploie que lorsque la tournure porterait trop certains lecteurs à la grivoiserie…

J’ai voulu permettre via un lien dans le menu de ce blog un accès rapide aux outils mis en place pour la recherche d’emploi : comme j’ai des accès quotidiens venant de Biblioemplois, j’en conclus que cela correspond à un usage réel (regardez dans la colonne de gauche : le lien existe à présent).

J’ai mis en place deux pypes pour la recherche d’emplois en bibliothèque, tous deux fusionnant les offres parues sur Bibliofrance et sur Biblioemplois :

1. une carte de France de ces offres d’emplois
   Utiliser l’outil / voir le billet explicatif
2. la possibilité de limiter l’affichage de ces offres à un ou plusieurs départements, avec abonnement RSS à la clé.
   Utiliser l’outil / voir le billet explicatif

Or, n’étant pas moi-même utilisateur de ces pypes, je n’en surveille pas l’efficacité.

Je m’aperçois au moment de proposer ces liens qu’aucun des deux pypes ne produit de résultat. Comme j’ai du mal à incriminer mon seul PC, j’en conclus que certains lecteurs sont passés par là, ont bloqué sur ces pypes bugués, et sont repartis…

Il me faut donc l’asséner ici : si vous tombez sur un billet vieux de plusieurs mois, mentionnant un pype (ou tout autre outil) qui vous semble d’un usage intéressant (même pour vous seul), et que vous constatez un dysfonctionnement, il faut absolument me le signaler !!!!

D’abord parce que mon honneur est en jeu, ensuite parce que c’est vraiment dommage pour vous et pour d’autres de laisser ces ruines en l’état, alors que je peux au moins essayer de les remettre debout.

Интересный сервис Yahoo! Pipes, через который можно обрабатывать контент с разных сайтов. Таким образом можно создать свою RSS-ленту или сделать автонаполняемый блог. Есть примеры и исходники.

http://pipes.yahoo.com/pipes/

Yahoo Spiel on Pipes:

“About Pipes

Pipes is a powerful composition tool to aggregate, manipulate, and mashup content from around the web.

Like Unix pipes, simple commands can be combined together to create output that meets your needs:

• combine many feeds into one, then sort, filter and translate it.
• geocode your favorite feeds and browse the items on an interactive map.
• power widgets/badges on your web site.
• grab the output of any Pipes as RSS, JSON, KML, and other formats. “

A few weeks back I decided to try out Yahoo Pipes and see if I could create a custom media feed about what was going on in my home town Gothenburg. Inputs for this custom feed was the Police’s RSS feed for daily situation reports and the RSS feed from three local newspaper, GP, GT and Metro. I decided to filter these feeds to only cover news articles about fires, the rescue service, police and the ongoing civil unrest. It only took a few minutes to construct the stream, and a bit of tweaking to let the correct items pass through the filter.

News and media are often used by the fire and rescue service to see what media is reporting, this is to get input about the situation and build situation awareness. In a discussion with Jonas Landgren we talked about how this tool could be used in ongoing or emerging situations. The idea was that a Yahoo pipe easily could be created to cover an event from a vast amount of media and official sources, thus eliminating the need required to manually locate news about the event. There might however be some problems with this approach. One problem is that you might lose news not picked up by the filter, and another problem is that Pipes can get quite slow when a lot of sources are added. Even though there might be a few problems I think that there could be a really interesting development in this area. If social networks more easily could be added, improved openness in systems used by response actors could be achieved and smarter filters were added this type of service would be even better. Custom  media streams could be a great resource for getting information and building awareness on the development, effects and emergence on a variety situations affecting the society.

Voir les billets précédents sur le Bouillon des bibliobsédés 2.0 :

• Les sources
• Le lien “Traduction” pour des billets en anglais

S’abonner au Bouillon (RSS)

Dans les 15-20 sources, il y a forcément des informations qui seront signalées plusieurs fois par les veilleurs.

Les objectifs étaient

1. de les dédoublonner
2. de faire valoir le nombre de fois où la même info était partagée (le taux de recommandation rend à peu près compte de l’importance, intelligence, pertinence de l’info)
3. d’indiquer, dans l’item dédoublonné, par qui il avait été partagé.
4. de conserver éventuellement les commentaires rédigés par les différents veilleurs.

Pour ce billet, on verra juste les deux premiers points

Module Unique (dédoublonnage)

Yahoo Pipes propose un module de dédoublonnage, sur un seul critère

Rappelons qu’à la source, on a une fusion de 17 flux RSS où les items ont été retriés par date de publication. S’y trouvent donc des billets, articles , etc. identiques, partagés plusieurs fois.

Cela constitue un gros fichier XML de 100 à 200 items, comportant chacun :

• un titre (item.title)
• une URL (item.link)
• une “description” (item.description) = contenu du billet
• une date de publication (item.pubDate)
• une “origine” (rajoutée à la source) = nom du veilleur

Le critère URL du billet (item.link) est le plus pertinent pour dédoublonner, évidemment.

Le module Unique fonctionne de la manière suivante :

• Pour chaque item, il prend en compte le lien.
• Il regarde dans la suite du fichier XML traité s’il y a des liens identiques.
• S’il en trouve, il les supprime.
• Dans le  1er item rencontré, le seul conservé, il stocke dans un nouveau champ item.y:repeatcount le nombre de fois où il a repéré le même lien.

Indication du nombre de recommandations

La valeur du champ item.y:repeatcount doit être ajoutée au titre, entre parenthèses.

Pour cela, on utilise une boucle (Loop) : pour chaque item, on concatène :

• le titre
• suivi d’une parenthèse “(” avec le signe “+”
• suivi de l’indication du nombre de fois où l’item aura été partagée
• suivi de la parenthèse fermante “)”

Et cette concaténation est mise comme nouveau titre

Ce qui donne :

Problèmes

Le module Unique ne conserve que le premier item rencontré.

Donc si, en dehors du critère “item.link”, il y avait des informations différentes dans les différents items dédoublonnés, celles stockées dans les items supprimés sont oubliées.

Par exemple :

Un veilleur utilise Google Reader (appelons-le Lully) pour partager sa veille, un autre utilise Delicious (appelons-le Couperin¹).
Ils partagent par hasard le même billet.
Si Lully a le premier mis ce billet en partage, celui-ci se retrouvera fusionné dans le pipe commun, avec dans le champ “Description” l’intégralité du billet.
Ensuite Couperin partage ce billet à son tour. Lors du dédoublonnage, c’est celui de Couperin qui sera conservé.
Or dans Delicious le champ “Description” n’est renseigné qu’avec la description éventuellement mise par le veilleur.
Donc dans le résultat final, l’item ne fournira plus le contenu du billet

Donc selon l’ordre dans lequel le même billet a été partagé, vous retrouverez son contenu dans le Bouillon… ou non.

Autre problème : dans l’item final, le seul champ “Origine” conservé sera celui du dernier veilleur, le billet eût-il été partagé 8 fois !

De même, si les 8 veilleurs ont rajouté un commentaire lors du partage, seul le plus récent sera conservé.

Bref, une perte d’information importante dans le cadre d’une veille collaborative (puisqu’elle rend compte et légitime le résultat final).

Donc les prochaines fois, on verra les plus gros morceaux : récupération de la liste des “origine”, et de la liste des “commentaires”. Sachant que j’ai aussi plusieurs autres billets à finir…

___________________________

1. Parce que c’est un autre compositeur de l’époque moderne.

Juste avant le lancement du Bouillon, a été rajouté un lien “Traduction automatique (Google)” qui n’apparaît que sur les articles en anglais.

Or les flux RSS des sources ne fournissent pas, pour chaque billet, l’indication de langue : cette information n’est fournie nulle part. Il faut donc :

1. demander au pipe de détecter automatiquement la langue
2. mettre la valeur de la langue (“en”, “fr”, etc.) dans un champ “language”
3. distinguer les billets dont la langue est “fr” de ceux dont c’est “en”, et pour ces derniers rajouter un lien.
4. Il faut que ce lien soit la concaténation de : “http://translate.google.com/translate?sl=en&tl=fr&u=” suivi du lien du billet. L’URL racine qui correspond à la traduction automatique dans Google, où la langue de départ est l’anglais (sl=en, pour start language) et la langue de destination est le français (tl=fr, pour translation language).

Etape 1 : détection automatique de la langue

<

Vous vous doutez que, lorsque vous lancez une requête dans Google, le moteur est capable de reconnaître la langue utilisée pour la requête. Donc Google sait reconnaître une langue sur quelques mots.

Il existe un service Google (une API, pour être exact) qui donne une valeur de langue quand on lui fournit une phrase.

Par exemple, la phrase suivante :

Comment utiliser la traduction de Google

peut être transmise à cette API ainsi :

http://ajax.googleapis.com/ajax/services/language/detect?q=Comment+utiliser+la+traduction+de+Google&v=1.0

Si vous cliquez sur le lien ci-dessus, vous ouvrez un fichier texte en format JSON ou les attributs de la valeur utilisée (la phrase en question) sont imbriquées les unes dans les autres, de manière assez comparable à un fichier XML.

Bref, dans la copie d’écran ci-dessus, le premier encart (Loop) contient un “URL Builder” qui crée un nouveau champ pour chaque item, champ intitulé item.lang_url. Sa valeur est la concaténation de :

• http://ajax.googleapis.com/ajax/services/language/detect?
• &q=item.title (titre de chaque billet partagé)
• &v=1.0

Le second encart (Loop) ouvre la page dont l’URL est celle qui vient d’être créée (avec un Fetch Data, dont l’adresse est item.lang_url).
Le champ “Path to item list” indique quelle valeur il faut récupérer, selon une syntaxe assez simple : dans le fichier dont l’URL est http://ajax.googleapis.com/ajax/services/language/detect?q=Comment+utiliser+la+traduction+de+Google&v=1.0, la langue est dans responseData > language. Donc on indique que la valeur responseData.language est à ranger dans item.language.

Ce faisant, Yahoo Pipes a récupéré en réalité toute une imbrication de l’information, et le dernier encart (Rename) simplifie le nom du champ : la valeur de langue (“fr”) avait été stockée dans item.language.0.content (on s’en rend compte assez vite dans le debugger). Elle sera désormais dans item.language.

Etape 2 : génération du lien vers Google Translate

Après que le champ item.language a été correctement renseigné grâce à l’étape 1, on divise le flux produit en deux. Le module “Split” produit deux flux identiques.

Dans celui de gauche, je ne laisse passer que les items de langue française, que je réinjecte dans le flux unique (Union).

Dans celui de droite, je bloque tous les items de langue française. Puis pour chacun de ces items, je génère la ligne suivante :

<p style=”text-align: right”> [créer un paragraphe aligné à droite]
<a href=”http://translate.google.com/translate?sl=en&tl=fr&u=item.link“>Traduction automatique (Google)
</a>
</p>

Puis je réinjecte ces items-là aussi dans le flux principal (module Union)

Et voilà.

Remarque

Notez que pour l’instant je n’ai pas abordé la question du dédoublonnage. C’est le plus gros morceau (rien à voir avec celui-ci). Donc j’avance progressivement (en tout cas, j’essaie !).

… collaboratif.

S’abonner au Bouillon collaboratif

Le Bibliobsédé a proposé une ouverture à la collaboration pour les ingrédients de son Bouillon. Vous retrouverez l’information chez lui et ailleurs.

Il me revient naturellement la joie de vous expliquer, aussi pédagogiquement que possible, le casse-tête qu’à signifié cette collaboration sous l’angle “Tutoriel Yahoo Pipes”.

Parce que, grâce à Silvère (que mille tweets chantent sa gloire !), j’ai dû me tordre les méninges et les boyaux et les tuyaux (de Yahoo) pour structurer un résultat correct, manipulable et intéressant.

Donc, présentation du pipe (et merci de le prononcer “païpe”, même intérieurement) en plusieurs étapes. Comme le résultat est relativement complexe, cela fera l’objet de plusieurs billets (au moins 3)

Aujourd’hui : Les sources

Chaque “veilleur” a un outil propre de partage d’informations. Google Reader pour la plupart, mais aussi Delicious ou d’autres outils encore. Chaque fois qu’un veilleur enrichit sa page de partage, il enrichit aussi le flux RSS correspondant. Donc l’activité de partage de chaque veilleur fournit un flux RSS.

La combinaison des sources n’est pas très complexe (ou n’est pas le plus complexe) :

• Un Fetch Feed pour intégrer le flux
• L’ajout d’un champ “Origine” pour indiquer à qui appartient ce flux (Module Loop avec un String Builder)
• Et la fusion de tout ça…
• retrié par ordre de date de publication immédiatement (pour voir aussitôt dans le debugger ce que ça donne)

Juste avant la production du fil RSS, on rajoute le champ “Origine” à la “description” (la “description”, pour un item dans un flux RSS, c’est tout simplement le contenu du texte)

Mais s’il n’avait fallu que ça, peut-être Yahoo Pipes n’aurait pas été nécessaire (d’autres outils existent pour fusionner des flux RSS !). Yahoo Pipes permet aussi :

• de dédoublonner
• de récupérer la liste des personnes qui ont partagé le même billet
• de récupérer d’éventuels commentaires de la part de ces personnes.
• proposer les billets partagés par plus de deux personnes unqiquement (parce que la veille de 20 veilleurs environ, ça peut paraître excessif à beaucoup. Ils peuvent alors profiter du Nectar du bouillon, par RSS ou par mail)

Cela, c’était le plus drôle.

Depuis, j’ai aussi finassé en rajoutant un lien “Traduction automatique Google” quand le billet partagé était en anglais. Donc il a fallu utiliser un outil de reconnaissance automatique de la langue.

Bref, vous voyez qu’il y a de quoi parler pour quelques autres billets !

Here’s a handy tool for y’all. WordPress. It’s probably one of the most common CMSs (Content Management Systems) in existence. It’s written in PHP and the MySQL dialect of SQL, two languages that I do not intend to cover in this course.  The beauty of WordPress, as opposed to systems like Ruby on Rails, Joomla or Django, is that you can do a lot of stuff without having to touch the PHP/SQL stuff. I might argue that the beauty of Django or Rails is that it makes the code so simple that you don’t mind messing with it. But that is definitely digressing. You might want to give those frameworks a look if you are going further later on.

The distinction with all these systems, the thing that makes them different from the javascript-based code we’ve been developing so far, is that javascript runs (typically) in the browser, on the computer of the user who visits your page. PHP and SQL though are executed on a the server. This is why UTS pays Dreamhost money for the usage of their machines.

But one step at a time. WordPress is a great way of seeing how all this website stuff works… let’s take it from where we left off last time, installing stuff from the dreamhost control panel.

The dreamhost add-new-software screen

Shazam. You now have wordpress installed. Why not take a look at what you have there, and fill out your info?

your brand new blog thingy - don't get the email wrong or you'll have to start again.

Once you’ve logged in you’ll be presented with a classic blog dashboard where you can hang around editing posts and so forth. There’s a million options here which might be confusing. Fortunately there is an extensive manual. Why not try to make a brand new post? Can you work out how to do that? Having edited your new post, why not find the post in a browser and have a look the published version? Now, run firebug, or press “view source” and have a look at the HTML that makes up the page… It’s a bit more complex that the stuff we were coding up by hand earlier, no?

where all the wordpress action is at

How is this magic accomplished? Well, if you’d like you can look deeper into the internals of the software… you can point Cyberduck at the server and have a look at the files that make up your wordpress installation. There’s a lot of stuff there- basically, a crapload of PHP code that automates the process of generating HTML for you. You could edit the PHP code in turn if you wanted to… but why bother? Life is short, and there are some easier options to try first.

FTP view of the php files that dreamhost has installed for you to make wordpress go

For a start, you can modify wordpress’s behaviour with plugins, little bits of PHP code that the community has generously written and shared. How about we look for ones that give the blog handy map mashup functionality? I happen to know one of the geo mashup standards is called “GeoRSS”, so I’ll do search for that. Plugins are variable in quality, but the first one listed there looks alright:

i search for georss, i find too many options.

From there it’s as simple as pressing the “install” button. Well, almost. this one uses google maps, so you have to sign on to google maps too:

signing on for a google maps API key is typical of the sorts of sign on you have to do to use third party services.

After those steps, BAM, you now have a mapping-enabled blog:

cripes, a whole new bonus settings menu has appeared.

Great. But you can also change the appearance of your blog. The visual equivalent of a plugin is a “theme”, a collection of CSS and code that changes the appearance of your blog. There are loads of community-contributed ones there too.

you can hunt "themes" the same way as hunting plugins

Of if you need to make finer changes, you can edit the CSS, or the code, that makes up your site by yourself:

The theme editor allows you to get at the CSS and HTML that makes up the site

Now, how about taking it further?

There are loads of interesting plugins – say if you want to integrate some other social networking sites that you are involved in, you could install  Lifestream

If you think giving your readers the ability to comment on the individual paragraphs of your blog, for example, there are plugins to enable that. There is an old system called comment press that makes an interesting case about why you might want to do that. Don’t install that, though, install the new version called digress.it.

Or if you wanted to use wordpress as a mashup engine, lots of things you might want to do don’t even need you to install plugins. For example, you could use the wordpress sidebar widgets to download and install new content from other sites (or from yahoo pipes, or from dapper) using rss… anyone care to walk us through that one?

And if you want to combine some of the other skills we’ve been learning with this… yes, jQuery, and CSS and all that stuff still work. jQuery ships with wordpress and it’s easy to enable. You can create scripts on a per-page basis or for every page in your site. You can also use the google-hosted version, as we’ve been doing in class. There’s a little trick, though – to enable the $() business to work like you’re used to, you have to write your javascript files like this:

jQuery(document).ready(function($) {
    // $() will work as an alias for jQuery() inside of this function
});

There are some excellent tutorials out there on precisely this:

• 30 Tutorials Combining Both WordPress and jQuery
• 15+ Plugins to Unleash the Invincible Power of jQuery and WordPress

品牌重整 企業識別設計更新及規劃… 紐約設計品牌專業顧問服務 Since 1983

商標(Logo)反映出企業的品牌形象，許多企業在成長到一個階段時，將商標加以更換或是修飾，以宣傳新的企業形象及價值，或者直接更換一個全新的商標；也有許多企業整合之後便更換商標，以針對有所改變的目標市場和利基客群,轉化或進增進對企業的品牌感受。重整或更新商標能提振品牌形象，使過時或逐漸萎靡的企業形象煥然一新。例如Inventec Electronic（英業達),  Twinhead International corporation（倫飛電腦), Stylution International Corporation (Restonic)（運時通集團/席樂頓), European Chamber of Commerce Taipei /ECCT (歐洲商會) PartnerTech Corporation（拍擋科技), Ingenuity Management Consulting Service / IMCI (智理顧問有限公司), Compass Public Relation（達豐公關顧問),  Taiwan Depository & Clearing Corporation /TDCC（台灣集中保管結算所)等等…以及其它的許多企業都藉由修飾 / 重整品牌形象的一連串動作，成功將事業推展至更高峰！

無論是新設計、修飾或重整企業商標，打造品牌形象無疑是對企業本身最重要的投資。正所謂：工欲善其事，必先利其器。因此，必須審慎尋找能夠對你的企業深入研究與了解的專業設計顧問…[…] http://newyorkdesignblogtalk.blogspot.com/2009/09/how-to-find-right-graphic-designer.html 才能攜手創造品牌奇蹟。

台灣集中保管結算所

倫飛電腦

智理顧問有限公司

瀧丰興業有限公司

運時通集團/席樂頓

歐洲商會

拍擋科技

全新視覺識別系統  拍檔科技股份有限公司正式啟用 -  紐約設計 受委託專業重整及整體規劃

拍檔科技股份有限公司 正宣佈式啟用全新視覺識別系統…紐約設計 New York Design專業重整及規劃…服務包括: Rebranding, Signature Makeover, Graphic Design, Problem Defining, Visual Identity Graphic Standards and Systems.

Partner Tech Corporation/ 拍檔科技股份有限公司: http://www.partner.com.tw

 Click Book Cover to view contents…

New York Design: http://www.newyorkdesign.com.tw

Since 1983, New York Design has been providing brand, corporate identity system; naming, corporate annual report, annual report online, packaging design, website graphic and marketing communication services to institutions and corporations, large and small.

平面設計,品牌策略,企業形象,命名,商標設計,品牌重整,識別系統, 網站設計, 整體品牌形象規劃…紐約設計 NEW YORK DESIGN…  since 1983

紐約設計 設計案例分享   New York Design Portfolio 2009

Click to view portfolio:

BRAND  AND PRODUCT NAMING PROGRAMS  -  商品,企業命名 –  運時通 美國運時通 席樂頓 美國席樂頓名床

A company name, a product name, a brand name are more than just a word or a name. It is the beginning of a conversation … the beginning to success… Our services from research, brand strategy, name development to recommendations. Over the years,  紐約設計 New York Design has had extensive experience in creating new names for companies both regionally and internationally.

Some of the companies for whom we have developed new names for include “TravelTalk” for American Express (TRS); “ArtMedia ” for Pacific Technology; “Taipei Metro” for Far Eastern Group (FEG); “Webiz” for QLifePlus Company ; “MobileTech” for MobileTech International Corp.; “FineSound” for Fine Sound Audio Systems; “SOTA” and “SotaPlus” for Twinhead Int’l Corporation; “ooolaalaa.com” for an US lady accessories retail chain… “MovieTIME” Apple iPhone home screen icon for Ooolaalaa-PCA Inc.

Project Demo :  運時通傢具集團 Stylution Group – 品牌命名,整體品牌形象規劃服務,企業識別,平面設計

Wichtiger Hinweis: Diese Idee bezieht sich auf selbst-gehostete WordPress-Installationen und nicht auf WordPress.com-Blogs!

Nutzt man WordPress als CMS, dann macht die Anzeige der eingehenden Links im Dashboard (Übersicht) keinen Sinn mehr, denn die stützt sich auf eine Google-Blog-Suche mit dem Paramter “link:” gefolgt von der eigenen Domain. Die Ausgabe erfolgt per RSS-Feed. Die Google-Blogsuche findet natürlich nur Blogs und gibt daher für die Sache nach einer normalen Webseite keine Ergebnisse aus. Man könnte das Widget jetzt einfach ausblenden, aber diese Funktion ist ja eigentlich ganz nett …

Was brauchen wir also? Eigentlich dasselbe nur mit einer normalen Google-Suche. Das Problem dabei ist, dass man die Google-Suchergebnisse nicht als RSS-Feed ausgeben kann. Nach einer kurzen Internet-Suche habe ich die Lösung in Yahoo Pipes gefunden. Hier einfach als als Suchbegriff “link:Domain-Name” eingeben und sich die Ergebnisse als RSS-Feed ausgeben. Fertig! Das einzige was man dazu braucht ist eine Yahoo-ID.

Verbesserungsbedarf gibt es allerdings bei der Anzeige in der Übersicht des Admin-Bereichs, denn dort steht nun folgendes:

Jemand verlinkt hierher in dem Artikel “Zitat-Auszug aus der Webseite”

Wobei das Wort “Artikel” ein Link zu dem besagten Artikel ist. Man sieht schon: In einem Blog gebe es einen Autor, der nun standardmäßig auf “Jemand” gesetzt ist. Sieht ein bisschen doof aus und verbraucht Platz. Das Ganze wird in der Datei /wp-admin/includes/dashboard.php definiert, aber so richtig mag ich mich an die Änderung von Core-Dateien nicht heranwagen. Ob es eine Möglichkeit gibt, das per Plugin zu lösen?

I tried adding my Twitter RSS feed to FeedBurner a little earlier.  Even though FeedValidator identified the feed as valid FeedBurner wouldn’t.  Googling this suggested that either Twitter isn’t producing nice RSS feeds (which FeedValidator’s validation seems to contradict) or FeedBurner has blocked it.

I decided to try passing the RSS feed through another tool.  I considered using Yahoo Pipes, but then decided that the feed aggregator from XFruits would be the quickest way.  I set-up a feed as an aggregation of just a single feed, passed that to FeedBurner and everything now seems to work.

an image from Kirsty Boyle's excellent girltron project

What on earth is a mashup? I don’t mean mashup in the media-sampling sense of DJ Earworm or Soda Jerk, but rather in the Web sense:
a mashup is a combination of two separate data sources available on the web, into some service or data set. Like the sampling sort of mashup, web mashups tend to be particularly valuable if their are surprising or subversive – but just plain useful is also good.

Here are some examples, in varying degrees of sophistication

• Got Gastro
• Essential Medicine Shortages
• Everyblock

Now, let’s look at the ingredients of mashups:

Have a poke around these guys, and see if you can propose some more mashups to me (hint- you can get started by simply picking your favourites off ProgrammableWeb)

So what are the ingredients of the mashup?

1. At least two data sets…
2. …in a recognised format…
3. …and some software to tie them together

1 some data sets

So, what are some datasets you can play with?

• some online databases – maybe too many, you’ll have to delve into these guys!
• a geographic dataset search
• This blog’s feed (can you remember how to find the address from last week?
• Hell, even the GotGastro data is online
• Anything with an RSS feed will do – how about your favourite blog?
• …or something locally relevant
• Want more? These, plus many more advanced datasets are on ProgrammableWeb

There are many, many more, and you can easily roll your own. Even this blog is a dataset.

2 a recognised format

Which brings us to the data formats question… Which I will gloss over for now. Let’s start by saying there are ways to convert between various formats, and some that are harder and some that are easier, and leave the detail until later. RSS, JSON, KML, should be no problem for any of us. Later on we can handle CSV, REST, and maybe more.

3 mashing software

This week’s practical exercise will be built using Yahoo Pipes, and will follow on from then Yahoo’s own examples page. Once we’ve had a play there, however, there are a plethora of other options that don’t even require any code – here are two excellent roundups, one from geek superstar Simon Willison and one from Sitepoint. There are lots of toolkits you can use to do this- google spreadsheets, google maps, xFruits (the only one I don’t regularly use), and dapper are among the most powerful. You might also be interested in looking at building mashups right inside your browser with firefox’s Ubiquity. There are also more complex tools for doing it all, like writing your own code, using maps and javascript…, or the mega YQL datasets

More advanced guides can be found, again, on ProgrammableWeb, or in one of the many guidebooks. (e.g.)

See if you can knock something together in-class. WARNING: things will break, go wrong and simply fail to work. Risks of the trade.

Finally, let’s be thinking about what all this enables. Let’s look at site with (maybe) valuable data that doesn’t enable mashups – the Australian government toilet locator. Can you think how this data might be more useful in combination with other sites? Has anyone else mapped their toilet data? (yes!)

Does anyone want to try to reproduce this gentleman’s amazing local goverment map?

Or: consider which data should be available to mashup. This is a hot topic in Australia right now, and there’s even money in it for you.

Se devo essere sincero, non so dire se si tratti di un caso isolato, fatto sta che mi sono trovato di fronte a dover risolvere un problemino della piattaforma Blogger. Ho deciso di iscrivermi tramite rss al blog di Innovatel. Tuttavia (in tutta sincerità) devo dire che non sono appassionato di ciclismo, pertanto quella parte di post non suscitava il mio interesse.

Ho deciso di sottoscrivere pertanto solo la categoria Linux, e nel caso non ci fossero state le categorie, il tag omonimo. In WordPress, fare una cosa del genere è semplicissimo, basta modificare l’indirizzo del blog in questo modo: http://andrealazzarotto.com/tag/linux/feed/

Né più né meno. Invece per Blogger non sono riuscito a fare una cosa del genere. Ho deciso quindi di utilizzare Yahoo! Pipes, uno strumento molto potente e versatile per manipolare i feed rss in modo miratissimo. Ha anche una interfaccia semplice (anche se all’inizio può confondere). Il problema però si è presentato qui. Non so per quale motivo, Pipes non riusciva a filtrare il feed correttamente.  Neppure il feed Feedburner del blog migliorava le cose (forse non era attivo l’ottimizzatore di compatibilità?).

Stanco di essere lì a tirare fuori box, cambiare parametri, collegare, scollegare, spostare… Ho deciso di cercare una alternativa. L’ho trovata in Feed Sifter!

Semplicissimo e senza registrazione, ho inserito le parole che obbligatoriamente ci dovevano essere nel post (ogni riga è un filtro indipendente, basta che uno sia soddisfatto). Così ora ho l’indirizzo del feed filtrato, che è questo:

http://feedsifter.com/?f=http%3A%2F%2Ffeeds.feedburner.com%2Finnoland&linux&open+source&software+libero&free+software

Sono contento! Posso seguire i post di Innovatel riguardanti l’argomento che preferisco, e tutto in modo semplice semplice.

Ah… ovviamente lo ringrazio perché ho usato il suo feed per fare questo articolo.