嗨 @anon094061，這是一個非常有趣的问题，因為它讓我們看到存在哪些不同的使用者和觀點。來自一位 IT 方面的「恐龍」的一些輸入。通常我討厭任何自動化的東西，除非它記錄得非常完善、透明且可配置（包括可關閉）。
災難是那些呈現概念混合的工具，您必須完成或確認某些工作或配置（通過 OK 或 ENTER），而其他視窗或部分則「沒有」任何內容，沒有視覺反饋，您只能希望您的設定已儲存或應用。或者也許您只是忽略了勾選符號，它不在您預期的右下角，而在右上角。聽起來熟悉嗎？
到目前為止，我在這方面對 PT 非常滿意，想不起來曾經在業務上失去過任何重要東西。我想是良好的習慣，或者良好的 PT，當然。
只有一個細節：對於本地團隊，我必須說，他們大多數從未使用過或見過那些「不太像軟盤」的東西。所以 PT 中的儲存圖示過時了，在這裡沒有幫助。
我在 PT 中工作了多年，甚至不知道自動儲存。有一天發現了切換章節時儲存的選項。PT 總是公平的，在切換章節時建議儲存，所以我們總是必須進行這些點擊。我理解這些訊息意味著「如果你切換章節而不儲存，你的工作將丟失」，所以當然我們通常希望在这种情况下儲存。
如果 PT 中除了切換章節的魔法之外還有更多「自動儲存」，我不知道。這裡的幾個帖子只是寫「自動儲存」。這具體指的是什麼功能？
否則，翻譯工作 - 或至少某些階段 - 可以通過輸入幾個選項來完成（我們通常在使用 / 符號分隔它們時，在旁邊使用幾個選項，作為團隊工作時），並在書面中和相鄰節經文的背景下查看它們。然後想法會被喊出來，螢幕上的文字會被編輯。這可以稱為「創意摸索直到變好」。
節經文歷史及其「時間標籤」對我來說仍然是一個深奧的謎。所以，如果應用於我們的創意測試方式，我們可能在一個節經文中可能有數十個或數百個版本，具有某些自動儲存功能，而不是可能十個版本：願它永遠不要發生。
所以，對於駕駛機動車或翻譯經文這樣的重要工作，我的觀點仍然是，機器，即使是 AI，也永遠無法知道真正發生什麼。所以使用者的角色是集中注意力並在需要時明智地儲存。或者使用選項離開並不儲存，如果某些草稿想法被考慮過但未被團隊喜愛。
我使用一些付費專業工具，如編輯器，有些讓我為復原和（自動）儲存目的分配特定數量的 RAM 或磁碟空間（以 GB 為單位）。這些是好功能。我個人可能在自動儲存方面非常保守（但我喜歡那些自動儲存被放入特定資料夾或對檔案名稱有特定後綴的工具，我可以自己配置）。相反，我通常會給復原非常寬鬆的額度，對於那些我確信它們真正處理好復原藝術的工具。
Hi @anon094061, very interesting question as it allow us to see what different users and perspectives exist. A little input from a dinosaur, IT-wise. Normally I hate anything automatic, unless it is very well documented, transparent and configurable (turn-offable included).
A desaster are those tools where they present a mix of concepts, where you have to complete or confirm certain work or configurations (with OK or ENTER) and other windows or parts there is “nothing” and no visual feedback and you just have to hope that your settings have been saved or applied. Or maybe you have just overlooked the tick-symbol which is not bottom-right where you expect it, but top-right. Sound familiar?
So far I am very happy with PT in this aspect and cannot remember ever having lost anything significant, workwise. Good habits, I suppose or good PT or course.
Just one detail: For the local team I have to say, that most of them have never used nor seen any of those “not-very floppy disks”. So the save-symbol in PT is outdated and not helpful here.
I was working for years in PT and never even knew about autosave. One day discovered the option to save when changing chapters. PT was always fair and proposed saving when changing chapters, so we always had to do those clicks. I understood those messages to mean “if you change chapters and do not save, your work would be lost”, so of course we mostly want to save in this case.
If there is more “autosave” in PT beyond the changing-chapters magic, I do not know. Several posts here just write “autosave”. What feature is meant by that exactly?
Otherwise translation work - or at least certain stages - can be done by entering several options (we often use several options next to each other with / symbols to separate them, when working as a team) and looking at them in writing and in context of the neighbouring verses. Then ideas will be shouted and the text on the screen is edited. This could be called “creative messing around until good”.
The verse-history and its “timing tags” is a deep mystery to me still. So instead of having maybe ten versions in a verse we might have dozens or hundreds with certain autosave features, if it were applied to our way of creative testing: may it never happen.
So my perspective for important work like driving a motor vehicle or translating scripture is still that a machine, even AI, can never know what is really happening. So it is the user’s role to concentrate and wisely save when needed. Or to use the option to go away and not save, if some draft ideas were considered and not-loved by the team.
I use some paid professional tools like editors and some let me assign a specific amount of RAM or disk space in GB for undo and for (auto)save purposes. Those are good features. I personally might be very conservative with autosave (but like it where those autosaves are put into specific folders or have specific suffixes to the file-names that I can configure myself). In contrast, I would normally give very generous allowances for undos, for those tools where I am confident that they really handle the art-of-undo well.
