"Ideal" desktop search: light version

The main problem facing a developer proposed last time systems for searching and processing information: many of its users will have to part with their habitual way of working with information. Well - here is drawn from the very problem the way to solve it. Below is a less risky (and hopefully less expensive) option for developers: light-version of the system, including a maximum of 20% previously proposed functions, oddly enough, gives you no less than 80% of benefits. And gives the key to "non-violent" introduction of a new style of work with information.

Version of light - this is a word processor (TP) with the functions, some of which proposed to use a full-version systems. Here's a short algorithm of a word processor.

F4 Step One
TP alone indexes on the user's computer text files of popular formats (*. doc, *. htm, *. pdf, *. txt, *. chm, etc.) - any use of Google Desktop index, or "personalized search Yandex" (contracts with their developers). Also, the program indexes the database of the most popular RSS-reader, e-mail client software for cataloging information, etc. (If the TA becomes popular, other third-party software developers themselves will begin to write for him converters).

Indexing - the key point of the program, allowing the user to save most of their habits. Our TAs, in fact, all the same, whether the user is looking for on Google and Yandex - or Quintura and Nigma; sorts whether the files in folders - and throws them into a pile; use in their projects if tens or hundreds of files with the facts - or heels -dozen svezhesohranennyh html-pages. Also, the user can not part with your favorite RSS-reader and other usual programs. It just adds to its arsenal easy to learn, fast and convenient tool for finishing information.

Step Two

Work begins TA traditionally - with the usual command "Open (Open)». Only in this case was called (more accurately - is pumping in a word processor) is not one, but a lot of files related to the information project. You do not have to look for these files on your hard disk - the program finds them itself. From you only need to ask a few key words, which, in your view, define the topic. But they should not ask a single line separated by commas - but as a tree.

For example, you specify two "parent" key phrases: "Automation of control" and "automation". And then to the first sentence add "child" keywords: "document», «ERP», «CRM», «BI», while the second phrase: "automatic lines", "machines", "machining center", " Robots "(TA will take into account the morphology). Press "OK" - and TA is pumping in the relevant branches of a tree query files containing specified keywords.

Something similar users see the same Google Desktop or "personal search Yandex" - in response to a query, these desktop search engines almost immediately issued a tape of the results. Only in this case we have no tape, and the tree results.

Important note: when the swap file in the program - a key to change on an unconscious level, style of work with information. Therefore, it should bring maximum pleasure, be fast and beautiful (for example, accompanied by the sound of chirping of lung machine for Money Counters and graceful unfolding result tree), and takzheizbavlennym of "glitches" and "hang".

Step Three

Tree results consistently unfolds in three vertical windows of the program: the sidebar, the editing window and the window of sources.

In this case, in the sidebar we see two key phrases linked to the "root" - the title of the project. This, as you remember the phrase: "Automation of control" and "automation". "Unfolding" them the same way as we are deploying a folder in Windows Explorer Windows, we will see the appropriate sentence for each child keywords. Further, when "deploying" a child of your keywords, we'll see the headers of files (or news RSS-posts), "podkachennyh" in this branch due to the fact that their texts contain relevant keywords.

Placing the cursor on any header, we will see in the second window of the program - ie editing window - the text of the file (RSS-message, email, etc.).

In Control Panel, TA, in addition to traditional text formatting buttons and text editing, there is a button-marker. If you select with the help of any text fragment (eg a paragraph or sentence containing the important facts, figures and quote a famous politician or businessman, etc.), then in the edit box will only have a dedicated marker segment. But the full text of the message appears in the third box - box springs. After processing the marker a few files and rearranging the cursor to the "daughter" or "parent" keyword or "root" of the project, we will see in the edit tape-marked by facts, each of which will be on its own gray base (all the others, the raw texts will be hidden in a tape as long as we and they do not allocate the needed facts - or until a cursor on the interchange header specific raw text. Such an arrangement, hopefully, reduce the risk of freezes the program due to lack of resources - while at the same time practically does not reduce the functionality of the program).

We can with your mouse or hotkeys «CTRL + arrow up / down / left / right" on the fly to change the structure and sequence of elements of the tree sidebar. We can also drag the facts relative to each other in the editing window, just picked up their mouse over the substrate.

I'll be back to their earlier description of the third window. So when the second window the cursor is located within the substrate of a fact in the third box, we see the full text of the original message, which earlier identified this fact-click the marker. In a more thoughtful reading of the original text in the third box, we can provide the same button and other facts - and they will appear in the second window, each in his gray substrate. All labeled for the facts in the third box, lying on a yellow backing. In Control Panel, TA button is present, an eraser, which allows to remove the selection from the facts (if they are acquitted of all - the original text go to the basket located in the sidebar, from which he, however, you can always get it back).

In Control Panel, TA is present the button of transition to a semi-automatic mode selection of the facts. When it is pressed, all texts will be automatically selected and placed on the gray substrate paragraphs, in which there are user-defined key words (by themselves and keywords in a semi-automatic, and manual mode is always illuminated). The user reads these paragraphs and estimates are interested in whether his presence in them of the facts (if not - removes selection). A small but important trick: if in one paragraph are present in the form of transfer of a few key words from the different branches - titles such communications sources highlighted in the sidebar in red. It is clear that the red headers will be duplicated in two or more branches of the sidebar. You can check these headers in the first place - and leave her there, where they are the same place.

The first three steps have been described so long to leave the smaller issues to developers. And now - the description for users.

TP loads in the texts on the subject, which you need to work out. - Read a text and highlights you need the facts. - On the fly shuffles the facts in the order that best reflect the theme. - Wrote the original article (analytical report), using as a "tracing" a logic circuit, the nodes which are selected and sorted your facts.

I'm not going to impose your opinion. Experience for yourself how the same problem is solved longer and harder in a traditional text editor.

In the first three steps described in the first version of the TA. If you find that the user accepts the major "pieces" of the program - the ability to automatically swap auxiliary materials, as well as the work is not a linear text, but the fact of the tree - it will be possible to complicate the algorithm TA.

In the second version of the program will not only simplify the finishing process information, but also to realize the tasks set in the previous article: "read, not reading" - ie use of information by studying it at the minimum necessary and sufficient level.

Instead, summary

In the second version of the TA will be implemented "semantic", "object" layers and the function of the appointment of key words synonyms - all that I described in detail in previous publications. If the developer will own experience in the field of semantic analysis of texts - they can also be used during "swap" of texts. Then the work will be structured as follows: the user defines the research topic one keyword and a set of synonyms for him, and the program automatically builds a tree of child concepts associated with the parent keyword, and playing in the texts of some important role. The user then manually removes the extra line and adds the right.

When users enter the taste of saving time when working with textual information - it was then, and they can be offered version of the full. I recall a super its development - creation of an effective conveyor searching and processing information based on the maximum reduction of routine operations and using a user action at a later stage.

The general conclusion: If your design requires a very abrupt change in habits of users - create an intermediate program. The main thing - properly define the functional unit from which actually begins the evolution of habits.

Feedback